FORTUNE — Last Wednesday, at a conference in Cambridge, Mass., Ben Bernanke sought to clarify the statements that shocked the markets just three weeks earlier. This time, the Federal Reserve Chairman reassured his vast, anxious audience that his pledge to start shrinking the Fed’s $85 billion in monthly purchases of long-term bonds, the latest version of “quantitative easing,” or QE3, didn’t mean that the Fed was abandoning the easy money policies that have cheered the markets for four years. The Fed would support the economy with “highly accommodative monetary policy,” intoned the Chairman, “for the foreseeable future.”
It was just what Wall Street wanted to hear. Over the next two days, the S&P (SPX) surged 1.6% to an all-time record close.
Bernanke was drawing a crucial distinction between the two, separate types of extraordinary monetary stimulus the Fed’s been deploying. The first is Bernanke’s long-standing commitment to hold the interest rates the Fed fully controls, those on short-term Treasuries, at exceptionally low levels. That’s the policy the Chairman promises to maintain until unemployment falls sharply, leading most observers to predict no change until 2015.
The second measure is the quantitative easing he now pledges to start winding down in later this year and end by mid-2014 if the economy keeps improving, as Bernanke expects.
MORE: The lowdown on China’s slowdown
The two monetary levers aim at different targets, spreading confusion over where interest rates are headed — the subject that now obsesses investors more than any other.
The first, extending virtually free overnight loans to banks, is a powerful force in restraining short-term rates, generally those on three-month to five-year Treasuries.
The second policy, purchasing gigantic volumes of long-term U.S. government and mortgage securities, is specifically designed to depress rates on bonds that mature much farther in the future, in 10 to 30 years.
MORE: Capital market climate change
The near-zero Fed funds rate that Bernanke will maintain does indeed restrain longer-term rates as well.
But its grip on 10- or 30-year yields is far weaker than the decisive power exerted by quantitative easing. Indeed, since May, just before Bernanke announced a probable end to QE3, the yield on 10-year Treasuries has jumped around almost one percentage point, to 2.6%, wiping out more than two years of interest payments. The markets clearly fear that far higher long-term rates are lurking in the absence of exceptional policies to rein them in.
That’s a crucial issue, because those rates are highly influential in determining the future performance of stocks, bonds, and real estate. Investors grant equities higher multiples when long-term rates are lower; both longer-maturity Treasuries and corporate bonds jump when rates decline; and developers pocket more cash flow from their projects when they borrow cheaply, raising the values of office and apartment buildings. When rates reverse course, so do all of those prices the Fed has been endeavoring to swell as a tonic for the economy.
MORE: Rising rates won’t end the stock market party
It’s important to understand why the Fed adopted quantitative easing, and why its inevitable end could prove risky. Between September of 2007 and January of 2009, Bernanke lowered the Fed Funds rate from 5.25% to less than 0.25% in an effort to quell the financial crisis. Longer-term rates also fell sharply, but not for long. By May of 2010, the yield on the 10-year Treasury had rebounded to around 4%. The Fed’s policies were clearly a lot more effective restraining rates on three-month and five-year Treasuries than those on bonds with longer maturities. So to boost asset prices, Bernanke initiated a policy specifically aimed at lowering yields on longer-term Treasuries, and all other bonds whose prices depend on those Treasuries, by initiating QE2 in November of 2010.
When the program began, speculation about a QE3 had already pushed down long rates, and the trend accelerated. From mid-2012 to mid-2013, the yield on the 10-year Treasury hovered between 1.7% and 2%, and the 30-year stood at around 3%, all exceptionally slender numbers. The policy also delivered just what the Fed wanted, an enduring bull market in both stocks and bonds.
Bernanke’s decision to shutter QE3, providing the economy keeps improving, has caused what’s called a severe “steepening of the yield curve.” That occurs when the “spread” between rates on Treasuries of different maturities, say the one-year versus 10-year bond, widen substantially. When that happens, investors are predicting that rates will rise rapidly from current levels in the years to come. In periods where the Fed isn’t deploying extraordinary measures, the yield curve is far flatter than today. “In normal markets, the spread between the one-year and 10-year is a little over 0.50 points, and add another 0.25 points to that for the yield on the 30-year,” says Chris Brightman, head of investment management at Research Affiliates, a firm that oversees strategies for $142 billion in funds.
MORE: Higher interest rates are a bitter pill for banks
With the recent surge in long-term yields, the spread between the one- and 10-year maturities has expanded to over 2.2 points (2.61% for the 10-year and 0.38% for the one year) — four times the normal range.
This steepening scenario doesn’t necessarily mean that stock, and especially bond, investors are facing big losses. The danger is that rates rise a lot faster, and a lot farther, than the increases the market is already forecasting. The chances that will happen are far from distant; we may have witnessed the start already.
To chart where we may be headed, let’s compare a typical yield curve to today’s severe tilt. In most periods, if one year Treasuries yield 4%, 10-year bonds return about 4.5%. Some people prefer buying one-year Treasuries, year after year, to purchasing 10-year bonds. Those folks are betting that yields on one-year bonds keep rising as time goes on. Hence, when they sell their bonds and replace them with new one-year maturities at the end of each December, say, they finish with an average return of at least the 4.5% over a full 10-year period. That’s just what they’d be guaranteed by purchasing the 10-year maturities today.
That’s the rationale for the yield curve: It projects that investors who buy a 10-year bond today will do just as well as those who purchase 10 one-year bonds over the same period.
So after five years, if those predictions are correct, the one-year bond yield would rise to around 4.5% and keep increasing to the 5% range over the next five years, giving the yearly buyer an average gain of that 4.5%. Once again, that matches what the investor in 10-year bonds received over the same period.
MORE: Ultra-low interest rates are making bonds unsafe
Let’s compare that gentle rise in yields to what the market expects today. The investor who’s now receiving 0.38% a year on the one-year bond, and who buys new bonds every year, will want average returns of 2.6% — equal to what the buyer of 10-year Treasuries receives — until 2023. So the yield on the one-year would need to rise to 2.6% or so after five years, then around 5% in a decade to generate that 2.6% average return. Sure, the jump is steep. But if that scenario occurs, the longer-term investor doesn’t get hurt. He or she still clips coupons of 2.6%, the average of what the person who buys new one-year bonds ever year receives, and gets their full principal returned on maturity.
The rub is that the Fed’s highly unconventional policies are heightening risk. Almost no one expected rates to explode over the past month. The average yield on 10-year treasuries over the past three decades is around 5.5%. Even today, the market expects rates to go back there, just not for around a decade. (In our example, if the one-year rate goes to 5% by 2023, it’s likely the 10-year yield will exceed 5%.) Let’s say it happens a lot more quickly. Investors in 10-year bonds will take an 8% loss for every one point that rates on new 10-year Treasuries rise, and owners of 30-year bonds will suffer 14% losses, or 42%, if comparable rates increase by three points, a fate that’s far from unforeseeable.
It’s impossible to predict what could make long-term rates shoot higher, now that the Fed is likely to relinquish most of its control over them. Most likely, it’s simply a shift in investor psychology, the realization that things always return to normal, and frequently in a hurry. “The market has come to conclude that long-term rates are abnormally low and will move up,” Martin Feldstein, the Harvard economist who served as chief economic advisor to President Reagan, told Fortune. “If short-term rates don’t stay low indefinitely, you don’t want to be the last person holding 10-year bonds at around 2.5%.”
MORE: A new Glass-Steagall won’t prevent another financial crisis
He adds that the Chinese won’t maintain their current appetite for U.S. Treasuries for long, and that the big looming deficits are another factor likely to pressure rates in the future.
The party that got us here was a bash. The return to normal could prove a long hangover.
