THE CONSENSUS THAT

a robust economic recovery is unfolding is now virtually unanimous. The world didn't end on Sept. 11.

flation is under control.

flation is under control. Even Japan is looking up. Maybe, as some upbeat commentators are saying, we never really had a recession to begin with. Happy days aren't here again...they never left!

Some of the most strident recovery voices belong to the same people who said "Don't fight the Fed" more than a year ago, and have been buying technology stocks all the way down. And now that they're in sight of breaking even, they're saying "I told you so." With all this happy-talk and false triumphalism, it's taken conviction to stick with the tactical asset allocation call I made in this column last Dec. 12 to sell the Nasdaq and buy long-term Treasurys. Those who followed my advice have been rewarded handsomely. Since I made it, the Nasdaq 100 has fallen by 10.8%, while the total return on long-term Treasurys has shown a loss of only 1.5%, despite a major back-up in rates.

This condition giving rise to my call remains very much in place, despite nagging from the recovery crowd about the imminent upside explosion in tech stocks. The reality is that the technology sector is nowhere near any kind of recovery that even begins to justify its still-lofty valuations. All the Nasdaq has to look forward to is Wyle E. Coyote's definition of a recovery: You jumped off a cliff, your rocket skates flamed out and you've splattered on the desert floor but at least you've stopped going down. To recap, the logic of my tactical asset-allocation call is based on value, risk and history. Based on today's consensus estimates for forward earnings, the earnings yield for technology stocks is far below the coupon yield of long-term bonds. Over history this "yield gap" has been about negative 1.1% on average. Today it is negative 3.2%, a level seen only two times previously: just before the crash of 1987, and at the top of the so-called bubble in early 2000.

This means that the structure of the U.S. capital markets is out of whack. One way to think about it is that bonds are paying you a premium to bear the risk that there will be a very rapid recovery. If there is a big recovery, as a bond holder you won't be hurt too much because yields have already risen substantially and if the recovery is less than rip-snorting, you'll get a windfall as yields come back down. But if you buy the Nasdaq, you pay the premium rather than receive it, and bear the risk that there won't be a rip-snorting recovery. If there is a big recovery you won't make much money because it's already priced in and if the recovery is mild, you'll get killed as multiples collapse. The market doesn't speak out of both sides of its mouth at the same time very often, or for very long. When it does, it's an opportunity and that's what this is.

But the siren song of the putative recovery can be very seductive. The lyrics are as follows: As a rapid recovery takes hold, overly pessimistic earnings estimates are going to have to be revised dramatically upward, and the yield gap will return to normal. The short version: This time it's different. The subtext: You might miss it!

But listen to what the CEOs and CFOs of virtually every big technology company have been saying at virtually every conference call this earnings season: We don't see any signs of recovery, and we can't guide beyond the next quarter. They mean it. That's not just code for "we're going to have a great back-end-loaded year," as many Wall Street analysts still seem to think, and many hopeful technology investors pray. Contrast this reality with the way technology stocks are valued now. Just take a look at Intel.

The average analyst forecast for 2002 puts earnings at 68 cents a share. With the exception of last year's disaster, those are the lowest earnings seen since 1995. And analysts aren't standing in line to revise them upward, either on Wednesday heavy hitters Jonathan Joseph of Salomon Smith Barney and Ashok Kumar of US Bancorp Piper Jaffray both cut their earnings and revenue forecasts.

But back in 1995, the last time EPS were lower than they are forecasted to be this year, Intel had a price/earnings ratio of 13. It was toeing the starting line for one of the greatest technology growth epochs in history. Today Intel's forward P/E ratio is 46. Does even the strongest ultra-"V"-never-had-a-recession-anyway recovery fanatic think that Intel is in a similar position today? I'm not trying to pick on Intel, even though I wrote about the company unfavorably in my column here two weeks ago. The fact is, all the other technology companies tell the same story.

Buying the Nasdaq here is an exacta bet. The recovery horse has to come in, and today's high valuations have to persist. Itchy tech investors worried that they might miss the move might as well admit they already have.

Investors don't like to listen to simple value arguments like this. But with bond yields as anomalously attractive as they are today in relation to the earnings yield of technology stocks, investors are going to find themselves swimming against the money-flow seeking the smarter risk-adjusted bet. Money at the margin will want to get paid for taking recovery risk in the bond market, not pay to take recovery risk in the Nasdaq.

Donald Luskin is chief investment officer of Trend Macrolytics, an economics consulting firm serving institutional investors. You may contact him at don@trendmacro.com.