It is clear from February's inflation data that there was a broad increase in price levels last month, especially for goods used during the early stages of production. The Producer Price Index rose 5.6 percent from its level a year earlier, up from 3.6 percent in January. On a month-to-month basis, the PPI rose 1.6 percent, doubling its recent pace. That increase was partially fueled by higher food prices, which makes up about a fifth of the overall PPI. Commodity prices tracked within the PPI Index rose 8 percent from a year ago, up from 5.6 percent last month.

Because of the price trends in food and energy, and the likely longer-term consequences of quantitative easing, investors and households are beginning to expect higher rates of inflation in the future. Implied inflation levels derived from the difference between 10-year nominal and real yields in the US have increased by 90 basis points since August. These recent yield levels suggest an annual inflation rate of about 2.5 percent.

Households are also expecting higher rates of inflation. Prices will rise by 4.6 percent over the next year, up from 2.2 percent in September, according to a poll by the University of Michigan. Over the next 5 to 10 years, inflation is expected to be 3.2 percent a year, versus 2.7 percent in September.

Those numbers - especially the longer-term inflation expectations - are not particularly high by historical standards. Nor are they high compared with the historical rate of inflation - which on a year-over-year basis has averaged 3.7 percent since 1950. This has left the Fed unfazed. In Chairman Bernanke's recent testimony to Congress he suggested that any rise in price levels from commodity inflation would be "temporary and relatively modest."

Another way to look at inflation expectations is to compare them with the trailing inflation rate over the most recent 12-month period. This is especially important for investors, because stock markets are sensitive to inflation rates rising quickly, even if the advance is from a low base (more on this below). Therefore, it's important to measure inflation and inflation expectations compared with recent trends in trailing inflation.

The graph below shows the difference in inflation expectations and the Cleveland Fed's median CPI rate. The Fed's median CPI tracks the "core" CPI well, but because it's not exposed to outliers it tends to be less volatile. The blue line in the graph below is based on year-ahead inflation expectations. The green line is based on inflation expectations over the next 5-10 years.

The graph shows that inflation expectations - relative to trailing inflation - have been generally well contained for 25 years. In the early 1980's, inflation expectations collapsed quickly, slightly outpacing the decline in trailing inflation (bond investors were more suspicious, as yields were much slower to decline during this period). But since the mid 1980's inflation expectations have generally been confined to about one percent above or below trailing measures of inflation.

Look at the recent spread. One-year expected inflation is now 3.6 percentage points above recent inflation. Longer-term inflation expectations are now 1.9 percentage points above recent inflation. Even if overall inflation expectations have not jumped to a level that concerns the Fed, it's clear that inflation expectations relative to trailing levels of inflation have broken out of their long-term channel.

It will be important to monitor this spread between inflation expectations and trailing inflation. If the spread stays constant then as the trailing inflation rate rises, inflation expectations will increase to levels where the central bank would take notice. Last month, New York Fed President Bill Dudley said, "If inflation expectations were to become unanchored because Federal Reserve policy makers failed to communicate clearly, this would be a self-inflicted wound that would make our pursuit of the dual mandate of full employment and price stability more difficult."

PPI Trend Levels

The topic of inflation tends to be a tool used by both sides of the debate about stock market performance. It's argued that because corporations can pass on rising prices of raw goods to consumers, earnings will keep pace with inflation, so equities are a good hedge against inflation. It's also argued that because the 1970's was a terrible decade to own stocks, very high rates of inflation must be bad for equities. As in many discussions surrounding financial market topics, there is some truth in each of these arguments. But the full story tends not to lend itself to such broad generalizations. As John Hussman observed a few weeks ago, stocks can benefit from inflation once it is widely anticipated and well-reflected in valuations, but otherwise, stocks are not a very good inflation hedge in periods when inflation is rising.

Maybe one of the most underrated risks regarding inflation is the speed at which it is rising, even if that increase is off of a low base. It's not high levels of inflation that precede important stock market declines, but instead how rapidly inflation is rising relative to its recent trend. And when you mix an overvalued market with rapidly rising inflation, bad outcomes tend to follow.

We can use the Producer Price Index to highlight this characteristic. The graph below shows each occurrence where inflation in the PPI Index was above 3 percent, the most recent PPI value was at least 60 percent above its 18-month moving average, and the cyclically-adjusted P/E ratio was above 16. For clarity, I've only displayed the first occurrence in any 12-month period. This set of conditions highlights periods where valuations were high (and therefore risk premiums are low) and producer inflation was rising at a fast pace relative to its recent trend.

It's clear from the chart that periods following high P/E multiples and quickly rising rates of inflation haven't worked out well for investors, on average. The worst instances came in December of 1965, three months prior to a 23 percent decline, in February 1973, a few weeks into a decline which would take stock markets down by half, in September 1987, and in September 1999 and October 2007, just prior to last decade's two 50 percent-plus declines. The May 1989 instance was early, as the economy didn't enter into a recession until the summer of 1990. But the market was mostly unchanged during this period, and would eventually fall by 20 percent. The one instance that was followed by gains came in July of 2003.

It's important to note that I'm not suggesting this as an investment model in itself. For one, it lacks a re-entry signal. It also leaves out important measures of investor sentiment and internal market action. Because multiples were low and inflation measures were flattening out, there was no signal prior to the nearly 30 percent decline during the summer of 1982, which marked the end of a 17-year secular bear market. That said, for a metric that use a single economic statistic and valuation criteria, it has historically highlighted important oncoming risks.

Note that the warning signal above doesn't demand high levels of inflation - just 3 percent, to indicate that some amount of inflation is already built into the economy. The most recent PPI inflation rate was 5.6 percent. All but one of the occurrence represented by the arrows above (the exception was in 2008) were at the same level or below the recent PPI inflation rate. The signals in 1966, 1968, 1987, and 2000 were all registered with a PPI under 3.5 percent.

The set of characteristics also doesn't demand high levels of valuation. The graph below puts this into perspective. Above each of the arrows for the above criteria, I've noted the cyclically-adjusted P/E ratio (which smoothes real earnings out over the prior decade). The market has run into trouble with far lower multiples than current levels of valuation when early-stage inflation was rising quickly.

Of course, the stock-market bubble era of the late 1990's through 2007 stands out. Valuations were higher at the peak in 2000 and again in 2007 than current levels, before stock prices declined. And the multiple was roughly the same prior to the 1966 decline. But the multiple was just 18 prior to the collapses in 1973 and 1987. Removing the valuation filter doesn't change the above sample set materially, it just shifts signals around slightly. That's because high levels of inflation relative to recent trends have often occurred in mature economic expansions, accompanied by high valuations - like those in 1973, 2001, and 2007 - just prior to the recessions that followed those signals.

In many ways QE2 has made the recent economic recovery much more divergent than most. Some metrics of this recovery resemble the characteristics of economy late in an expansion - including measures of production activity and increases in producer price indexes - while some components still resemble the characteristics of early-stage recoveries - like the number of net new jobs being added to the economy. When the PMI Index has been above 60 (it's currently 61.4), the average unemployment rate has been 5.5 percent (it's currently 8.9 percent). Price indexes are similarly stretched. When the PMI Price Index has been above 80 (it's currently 82), the jobless rate has typically been 5.2 percent. So QE2 has produced a more "stagflationary" economic profile than is usually the case.

Higher prices are also increasing the pressures on households whose incomes continue to stagnate and who continue to struggle with debt loads. The imbalances within this recovery are likely going to represent the next obstacles to higher valuations - or maintaining current levels of elevated valuations - that investors confront. The divergences in the characteristics of this recovery will likely become even more visible as higher prices of raw goods work their way into measures of consumer inflation.

Consumer Inflation

The graph below applies nearly identical metrics to the slightly less volatile CPI Index. The arrows on this graph show where the year-over-year change in CPI was above 3%, the inflation level was 50 percent above its 18-month moving average, and the P/E ratio was above 16. Applying these criteria to changes in the CPI Index gives fewer signals - no signal was given prior to the 1969/1970 decline, the 1991 decline, nor prior to last year's correction. Still, tracking CPI momentum has proved valuable by signaling what we can justifiably call the Four Big Ones: 1973, 1987, 2000, and 2007.

This time in the graph, I've plotted the levels of the year-over-year change in the CPI at the time of each signal. In this case, these are more instructive than the signals themselves. Investors don't wait until inflation is rising at 15 percent a year to start aggressively selling overvalued equities. They don't wait for 10 percent or 7 percent inflation, either. The highest level of inflation within this group of signals was 5.5 percent, prior to the bear market of 1973-1974. All of the other signals that preceded declines coincided with rates of consumer inflation somewhere between 3.2 and 4.3 percent. It's the combination of high equity valuations and the acceleration in the rate of inflation - even if that increase is off of low levels - that has represented the most dangerous periods for risk taking.

In February, the CPI was up 2.1 percent from a year ago. This was up from an annual inflation rate of 1.6 percent in January. It's well within reason to expect the rate of consumer inflation to rise above 3 percent over the next few months. A continuation of the recent trend would put year-over-year inflation above 3 percent by May. Keep in mind that the CPI Index hit its low in June 2010. So the year-over-year calculation will get a natural boost as the index moves higher over the next few months. If the CPI climbs above 3 percent by June, it will also almost certainly be more than 50 percent above its 18-month moving average. The CAPE Ratio is currently 24.

The start of a contraction in valuation multiples has historically come long before high levels of inflation. Rather, it's inflation rising faster than its recent trend - sparking concerns of higher longer-term inflation - that increases risk aversion among investors. When these periods occur alongside high P/E ratios, the contraction in multiples has usually been substantial. Ben Bernanke recently told Congress that he is confident that the recent rise in the rate of inflation will be temporary, and that inflation expectations are unlikely to become unhinged. Stock investors must be able to share that belief and that forecast, because a change in longer-term inflation expectations - even from a low base - would increase stock market risks importantly.