This series, brought to you by Yahoo! Finance, looks at which upgrades and downgrades make sense and which ones investors should act on. Today, our headlines feature new buy ratings for Microsoft (NASDAQ:MSFT) and Allison Transmission (NYSE:ALSN). On the other hand, Alcoa (NYSE:AA)...

"... should trade at a discount"
That's the conclusion that Nomura Securities came to this morning, at least. The aluminum giant reported Q4 and full-year earnings yesterday, and while management touted its results as delivering "strong operating performance," Alcoa nonetheless missed  analyst targets by more than 30%, reporting only $0.04 per share earned in the quarter.

Quoted on this morning, Nomura is seen praising Alcoa for showing "positive momentum in the downstream segments," which the analyst calls "highly valuable." In the company's Q4 report, Alcoa pointed out that it's now getting 80% of its operating profits from downstream, value-added aluminum products, despite these accounting for only 57% of its revenues.

Yet even so, Nomura notes that these valuable assets "are unlikely to be spun off from the parent." Nomura argues that Alcoa should not receive full credit for the "sum of the parts" of these operations, but rather should be dinged with a SOTP discount for leaving them trapped within a less profitable upstream aluminum superstructure. Predicting that the company will earn only $0.28 per share in fiscal 2014, the Nomura continues to rate Alcoa only neutral -- and says the shares are really only worth about $8 -- 20% below where they trade today.

I agree. In fact, I'd go even farther. Given that Alcoa is currently unprofitable, but assuming that Nomura is right about this year's earnings coming in at $0.28, this works out to a price-to-earnings ratio of about 36 on the stock. This might seem a fair price to pay, since analysts predict the stock will grow its earnings at 38% annually over the next five years. But there are two problems with this thesis:

First: 38% annual earnings growth? For a smokestack industry aluminum smelter? Not likely.

And second: If you factor Alcoa's debt into the picture, the company's enterprise value is currently 49 times this year's projected earnings. Even if you buy the 38% growth argument (and I don't), that's clearly too high a price to pay.

Transmission read: "Buy!"
Sticking with the smokestack industries a while longer, we turn now to Allison Transmission, which this morning received a new buy rating from Longbow Research. Priced at 38 times earnings, Allison actually costs a bit more than does Alcoa. Yet according to Longbow, this $27 stock is likely to rise to $33 per share over the next 12 months. Is it right?

Maybe, but more likely not. On the one hand, at 38 times earnings, but with projected earnings growth of less than 9%, Allison looks like a pretty expensive stock. What these numbers don't tell you, though, is that over the past year, Allison has generated more than two-and-a-half times as much real free cash flow ($356 million) as it has reported as GAAP net income ($134 million). As a result, with a price-to-free cash flow ratio of just 14, Allison is not nearly as overvalued as its P/E ratio makes it appear.

The problem, though, is that 14 times FCF is still too much to pay for a sub-9% grower. Plus, the company carries a substantial $2.5 billion net debt load. Factor that into the valuation picture, and I'm afraid Allison stock still looks quite overpriced -- and Allison's chances of returning the 22% profit that Longbow is promising, exceedingly slim.

Time for Mr. Softy to shine?
And finally, today, let's end on a more optimistic note -- about Microsoft. This morning, Brit banker Barclays announced it has decided to upgrade Microsoft to overweight. Its theory being that investors have overreacted to Alan Mulally's announcement that he won't be abandoning Ford to ride to Microsoft's rescue -- and are missing the story of Microsoft's improving business prospects elsewhere.

While admitting that Microsoft faces some headwinds from a PC industry in decline, Barclays is optimistic about several trends at Microsoft. Citing reports from market researcher Gartner that "cloud services" will grow into a $200 billion annual market over the next two years, Barclays points out that between Azure and Office 365, Microsoft is now raking in about $2 billion a year in cloud revenue, with Azure in particular growing at a 100% annual clip.

Barclays also likes the growth prospects for Surface touchscreen computers and Lumia smartphones. According to the banker, other analyst expectations for these businesses are "now so low that upside could start to show in the coming quarters, while Windows comps are also getting easier."

Personally, I don't know whether all of these predictions will pan out as planned. But I have to say that Microsoft's valuation looks enticing. Priced at just 13.4 times earnings today, Microsoft is so flush with cash that, excluding cash and debt, its P/E drops to just 10.4. Meanwhile, free cash flow continues to exceed reported net income, with the result that if you value the company on its enterprise value to FCF, the ratio drops even further, to just 9.9.

Call me an incurable optimist, but if Microsoft just hits consensus growth targets of 7% and maintains its current dividend payout of 3.1%, this is enough to make the stock fairly valued at worst, and a bargain at best.

Fool contributor Rich Smith has no position in any stocks mentioned. The Motley Fool recommends and owns shares of Ford. It also owns shares of Microsoft.