When things looked depressingly bleak in early 2009, I started inching into corporate bonds and found that I could buy bonds of excellent companies for 25-39 cents on the dollar. The idea was: Why buy the stock when you can get all of this upside with the bonds and get paid 25% while you are waiting? The "bond fire sale" closed down pretty quickly and, as I redeployed into stocks, I constantly tried to compare the relative merits of individual stocks and the stock market in general to bonds. It is a complex, difficult and ultimately arbitrary exercise, but I think that every investor has to think it through - especially now.

In the 1990's, the "Fed Model" was in vogue; it involved comparing the yield on 10-year Treasuries with the earnings yield (the inverse of the price earnings ratio) on the S&P 500. If 10 year Treasuries were yielding 5%, this would suggest that the earnings yield on the S&P 500 should also be 5% and, thus, that the price earnings ratio should be 20. I have written a piece on where these kind of calculations would lead in today's market - the result is a PE of over 50 and a Dow 30 of over 40,000. Recognizing that these results seem absurd, I have tried to revise the model to produce a more realistic comparison of stocks and bonds.

First of all, I think that corporate bonds are a better metric than Treasuries. After all, it doesn't do corporate America much good if the government can borrow at low rates, but corporations face tight credit - as they did in early 2009. Secondly, I prefer 5 year rates to 10 year rates because it is simply too difficult to look out 10 years and see anything more distinct than fog. It is also the case that, for many purposes in the real world, corporations borrow for roughly 5 year time durations.

I also believe that, in a world starved for yield, income delivered to shareholders as dividends should be treated differently from income retained by the corporation. I think investors are beginning to put a higher value on dividends, and, more importantly, dividends are more directly comparable to interest payments on bonds. An investor gets the dividend check in the mail and can invest it or spend it as he or she pleases. For example, if Apple (NASDAQ:AAPL) sent me a check every quarter, I would not deploy the cash the way AAPL does (short term Treasuries and securities yielding less than 1%) and, frankly, I would probably rather get 80 cents per share in the mail and invest it myself, then have AAPL sit on one dollar per share the way it does.

So, we wind up with a complex model which is still somewhat of a work in progress. Because some of the numbers are the product of arbitrary assumptions, readers should feel free to substitute other numbers that better reflect their own opinions and levels of risk tolerance. What is important is to THINK about the stock/bond comparison and to attempt to do it in a systematic way.

The bond yield that I am using as a kind of benchmark is the yield on 5 year BB bonds. I think this reflects corporate borrowing costs for many mid-sized companies and it also gives us a set of bonds for which there is some default risk. Although this yield has jumped around and I have seen it calculated various ways, I am going to use 4% (roughly 320 basis points above Treasuries) as a metric for our calculations. Another way to look at it is that these bonds are trading at a 25-to-1 price-earnings ratios.

Lets use a stock I follow quite a bit - Microsoft (NASDAQ:MSFT) - which closed Friday at 30.24. MSFT pays a dividend of 80 cents a share and this year's earnings are projected to be $2.76. So, it pays out 80 cents a share of earnings as dividends and it retains $1.96 of earnings. As to the dividends, there is a pretty good case that they are safe and are much more likely to increase than decrease over the next 5 years. But I will attach a 100 basis point risk penalty to them and essentially require a yield of 5% on the dividend part of the earnings. Thus, as to the dividends, the value of the company is 20 times 80 cents or $16.

As to the retained earnings, there is certainly more risk that: 1. They will fluctuate (dividends are not likely to be reduced if there is a short term reduction in earnings) and 2. They will not redound to the benefit of the shareholders. For this reason, I will attach a 300 basis point penalty on the retained earnings segment of the total earnings - requiring a 7% yield. Multiplying the $1.96 of retained earnings times 14.3 produces a value of $27.03. Thus, the value of MSFT as a business operation is $43.03. But MSFT also has $5.70 a share in net balance sheet cash; thus, the value of the company as a whole is $48.73 - or about 60% north of where we are.

You may disagree with one or more of the assumptions I made, or the data I used, and I invite readers to substitute various durations and bond ratings and make calculations of their own. There is a case that we should use MSFT's "expected" dividends and earnings over the next 5 years to make the comparison fair, and this would push numbers up quite a bit. On the other hand, the spreads that I used to discount the value of MSFT's dividends and earnings may be too generous. There are an infinite variety of ways to make this comparison. It is also the case that this methodology tells us nothing about the comparative merits of yield vehicles such as agency mortgage REITs - like Annaly Capital (NYSE:NLY) - and Business Development Companies - like Ares Capital (NASDAQ:ARCC).

But as we experiment with different spreads and different assumptions of future growth, one thing is becoming very clear: it is getting harder and harder to come up with a set of assumptions that makes bonds look attractive in comparison with dividend stocks.

**Disclosure: **I am long AAPL, MSFT, [ARCC, NLY.