For the second day in a row, the economic data fairy left a lump of coal in the market's stocking. January New Home Sales, expected to show a bounce-back from last month's decline, instead fell significantly further. The 11% decline took the annual run-rate to just 309,000 units. That's not just disappointing, it's an all-time record low (see Chart, source Bloomberg, below).

Not just low, but the lowest ever.

The presence of massive government stimulus always makes it difficult to tell what is happening, organically, beneath the stimulus. Late last year, when home sales were bouncing, however feebly, it was not easy to discern how much of that was true organic growth - a turn in housing market fortunes perhaps sparked by the government tax credits - and how much was merely pulling demand forward. Sometimes, it's impossible to ever tell: if the bottoming in New Home Sales happened to coincide with the onset of government stimulus, there would be no way to tell how many New Home Sales were "created or saved" by the stimulus compared to the natural growth that would have happened anyway simply because the economy was enjoying the upside of a natural oscillation.

It now appears, though, as if there was never a turn in the market after all. It appears, right now anyway, as if a large proportion of the bounce was in fact due to the pulling of demand forward from 2010 to 2009. Certainly, it is difficult to make the argument that we're better off having spent the money to goose the housing market, since it didn't stay goosed.

I am sure that some observers will blame the weather in January. Perhaps that is some of the underperformance, since the biggest decline in sales was in the Northeast. But I seem to recall that the weather in January is usually pretty poor and the seasonal adjustment provides for much of that.

Now, the good news here is that most home sales don't happen in January, so a disappointing seasonally adjusted run rate in January is not as bad as a disappointing rate of sales in late springtime. But it's still bad news.

Equity market participants, however, seemed not to care much. Stocks came into the day on fire and, after briefly retracing some of the gains on the New Home Sales data, launched higher to end the day +1% while bonds languished (a weak sale of 5y notes didn't help. Sometimes $42bln really feels like $42bln!).

So stocks didn't care about weak home sales or, more likely, they cared more about Chairman Bernanke's reassurance that the Fed isn't going to be raising any relevant market rate (whether the Fed Funds rate or the Interest On Balances) any time soon. Although he merely repeated the "extended period" language, stock investors breathed a sign of relief because, after all, lower rates are good for stocks.

When buyers of residual claims on the earnings of corporate entities (that is, stocks) begin to look to the level of interest rates as the reason that stocks are a good buy, it is starting to look like the last refuge. Loving stocks because interest rates are low is a really weak reason.

"But wait!" some will cry, "what about the 'Fed model'?"

The so-called "Fed model" proposes that there is a relationship between equity values and interest rates since like any other capital asset, equities can be appraised as the present value of future cash flows, and present value calculations depend rather heavily on the rate of interest used to discount the flows (e.g., the Gordon growth model is a spare model that declares the value of a stock to be equal to the current dividend discounted by the difference between the required return r and the dividend's assumed growth rate g; obviously, the required return is related to other returns available in the market). So, when interest rates are high, equity multiples "should" be low; when interest rates are low, equities "should" be high.

But this confuses the problem of explaining the level of P/E multiples with the problem of deciding whether such a P/E multiple is an appropriate level - that is, whether stock values are priced to return an adequate return to compensate for the risk. The "Fed Model" does a passable job of explaining why high P/E multiples might be associated with low interest rates, but it tells us nothing about what happens next. A fantastic article by a man far, far smarter than me is "Fight the Fed Model" by Cliff Asness, which appeared in the Journal of Portfolio Management in 2003. Mr. Asness argues persuasively (and with his usual biting wit) that (a) the Fed Model explains a historical relationship, (b) the rationale given for the Fed model (one version of which I just provided) is wrong, which means that the historical relationship just illustrates a systematic investor error, and (c) there are much better ways to predict the likely future return to stocks, for example the simple level of P/Es themselves. It turns out that historically, the higher the starting market P/E, the lower the subsequent return.

Mr. Asness is right, but I don't even think you need to get to perfect theory to see that relying on the level of interest rates as a reason to buy stocks is faulty logic. When interest rates are low, and therefore equity multiples are high, what comes next? As the old Marty Zweig commercial says, "if you can spot meaningful changes...in interest rates and momentum...you'll be mostly in stocks during major advances and out during major declines. [emphasis added]" Rates are low, and the Fed will keep them low as long as they can. But what's next? Are they going lower? Short rates simply can't go any lower, and if long rates go lower it's probably indicative that we are in a depression. No, the next major move in interest rates is far more likely to be to higher rates than to lower rates.

According to Mr. Zweig, that means we need to be mostly out of stocks.

Personally, I don't have a strong opinion that stocks need to go drastically lower in nominal space. With a decent rise in inflation, they could "correct" to much lower real values while chopping sideways for a few years (which is what happened in the 1970s of course). But again, such thoughts are not much solace.

On Thursday, Durable Goods (Consensus +1.5%/+1.0% ex-transportation) is expected to follow a positive print with a positive print. Many years ago, I examined whether I could beat the consensus by always forecasting (-0.5 times last month), so if last month was +2%, my forecast would be -1%. At the time, it turns out that such a strategy actually did beat the consensus by a small margin, which mainly goes to say that Durable Goods are extremely volatile and it's hard to make much out of them. That doesn't change the fact that the market enjoys reacting to Durables.

Also due out is Initial Claims (Consensus: 460k, after the surprise 473k last month). Optimism reigns supreme as Claims keep exceeding the Consensus week after week. At some point, forecasters are going to have to adjust, or the numbers are going to have to start improving. Which is more stubborn, economists or the economy?

Of course, the last piece of news is that another storm is supposed to hit the New England/New York area and bring a lot of snow. That probably won't have a lot of impact on trading early on, but expect conditions to thin somewhat as the day goes on.