Valuation can be an inexact science, with multiple people often getting different numbers for the same company. It could simply be the difference between one's expectation for anticipated growth, or possibly some qualitative factor that is weighed more heavily by one person over the other. With so many different ways to look at where a company might be going, it can get a little overwhelming at times if you try to do something yourself.

One quick valuation tool that I favor is the Graham number. I discussed this number in depth last year, and every so often, I like to take a look at the stocks of the S&P 500 to see if there are any great values based on Graham number valuations. Luckily for us, Benjamin Graham gave us a quick shortcut to help us determine whether or not a stock is considered "cheap": In an ideal situation, the P/E ratio and P/B ratio multiplied together should not exceed 22.5, with a maximum P/E of 15 and P/B of 1.5. It is then that we can use the Graham number to calculate potential upside for any number of stocks.

The quick test narrows the field
In my latest examination of the S&P 500, an interesting trend emerged. While looking for companies that met Graham's quick test above, I was left with 47 companies from various industries and sectors. The surprising thing about this list, however, was the prevalence of companies from the financial sector.

More than half of the companies -- 27 in total -- came from this sector, helping to show that there might be a lot of value in this often beaten-down sector. When I actually calculated the Graham numbers for all 47 companies, financial companies took center stage once again.

Financials lead the way!
The formula for the Graham number is pretty straightforward: Multiply earnings per share by book value per share, then multiply that by 22.5, and finally take the square root. The result, in dollars, is the Graham number. Of the 47 companies that passed the quick test, only one -- consumer goods maker Mondelez International (MDLZ 1.91%) -- currently trades above its Graham number valuation. The remaining 46 companies all have room to grow into their current Graham number, but the five with the most upside all came from the financial sector, with four of the five being insurance companies:

Company

EPS (TTM)

Book Value Per Share (MRQ)

Graham Number

Recent Price

Upside

Genworth Financial (GNW 0.33%)

$0.65

$33.62

$22.17

$8.26

168.5%

Lincoln National Corp (LNC 2.21%)

$4.56

$54.71

$74.92

$28.49

163%

Assurant (AIZ 0.44%)

$5.67

$64.84

$90.95

$40.65

123.7%

SunTrust Banks (STI)

$3.59

$37.87

$55.31

$27.40

101.9%

Unum Group (UNM 0.86%)

$3.17

$31.87

$47.68

$23.80

100.3%

Source: Yahoo! Finance and author's calculations; TTM: trailing 12 months; MRQ: most recent quarter.

Why so cheap?
Any time a company trades at a dramatic discount, it is important to look at the reasons. Genworth Financial, for example, has fallen recently as it tries to divest itself of some poor performing segments. It also joined Lincoln National in running into some problems last fall as the state of New York began investigating the practice of captive insurance and managing risk. Assurant felt the wrath of Hurricane Sandy last October and has seen a (hopefully) short-term impact from claims arising from that storm.

Foolish bottom line
Because my primary focus tends to be in the financial sector, over the next few weeks, I will take a closer look at the 27 financial companies that are "cheap" according to Graham and determine if they have potential beyond a cheap valuation. As in the case of Assurant, it could simply be short-term forces bringing down the price, or it could be something that could linger with the company for the long term.

While I will focus primarily on financials, that doesn't mean that you can't look at the other 19 companies in the S&P 500 that are currently trading at a discount to their Graham numbers. I personally think the Graham number is a useful jumping-off point when first looking at a stock, and it doesn't require a whole lot of effort to calculate.