Edging closer to the original low PB, high F_Score formulation

Having modified the Nifty screen last week, I've had another look at my Thrifty screen today.

While The Nifty screen targets good companies at cheap prices, the Thrifty screen targets viable companies at bargain prices. To differentiate them, and discover very cheap stocks, I've decided to rely more heavily on the F_Score as a measure of viability, and toughen up the measure of value.

Before I used price to book value, the market value of the company compared to its value according to accountants, as the value factor, now I'm switching to price to tangible book value. I'm counting only a selection of the company's assets: property, machinery, vehicles, stock, cash, and receivables the company owns, and excluding goodwill and other intangibles.

While I generally work on the principle that book value is as good a measure of the value of a company's assets net of liabilities as any I could come up with, in the case of distressed companies trading at low valuations I'm not sure sure. Often these companies are restructuring extensively, and they may still be writing off the value of their assets. Prime candidates for write offs are intangible assets, especially goodwill, when companies discover investments they made in the past are not living up to expectations and are worth less than they thought. It may be prudent to ignore intangibles altogether, and that's what I'm going to do, for a while at least.

For companies to be viable, they have to not only own physical assets we can put a value on, but they must earn a profit with those assets, if not in their latest results, then soon. Since the F_Score compares how a company is doing compared to the previous year across a broad range of measures, it's good tool for discovering recovering companies. However I also want some evidence the company has been profitable in the past, which is why only companies that have made a profit over the last ten years are included.

The Thrifty screens are ranked by these two variables, PTBV (the lower the better) and F_Score (the higher the better), alone. I used to add two more variables into the mix, the 10 year average return on equity (higher the better), and leverage (the lower the better). I think they may be muddying the water, and since both are also variables in the Nifty screen, they direct the two screens to very similar companies.

Distressed companies are more likely to be in debt, and their past earnings may be less relevant to the future, so by screening out companies with high levels of leverage and lacklustre profitability I may be screening out genuine bargains. Debt and poor profitability could be the reason the shares are so cheap, in which case it's the company's ability to repay debt and improve profitability that matters, factors indicated by the F_Score.

Finally, and with great trepidation, I've removed the requirement that companies must have reported their full year results in the four months preceding the screen. The danger now is that some of the data is old. Companies can take months after their financial year ends to report their results, which means in some cases the data is from periods that started two years or more ago.

A lot could have happened since then, and I'll have to consider those events when investigating companies, to be confident they don't undermine the statistics.

The advantage of removing the date restriction is a lot more companies enter the screen, a particular problem at this time of year when relatively few companies are reporting results. Some of those don't deserve to be there, but that's a risk I'm going to take.

Here's the table, only the top 10% of companies meeting the criteria are shown:

Here's the spreadsheet.