The Quant Result


I like this.


An eternal optimist, Liu-Yue built two social enterprises to help make the world a better place. Liu-Yue co-founded Oxstones Investment Club a searchable content platform and business tools for knowledge sharing and financial education. also provides investors with direct access to U.S. commercial real estate opportunities and other alternative investments. In addition, Liu-Yue also co-founded Cute Brands a cause-oriented character brand management and brand licensing company that creates social awareness on global issues and societal challenges through character creations. Prior to his entrepreneurial endeavors, Liu-Yue worked as an Executive Associate at M&T Bank in the Structured Real Estate Finance Group where he worked with senior management on multiple bank-wide risk management projects. He also had a dual role as a commercial banker advising UHNWIs and family offices on investments, credit, and banking needs while focused on residential CRE, infrastructure development, and affordable housing projects. Prior to M&T, he held a number of positions in Latin American equities and bonds investment groups at SBC Warburg Dillon Read (Swiss Bank), OFFITBANK (the wealth management division of Wachovia Bank), and in small cap equities at Steinberg Priest Capital Management (family office). Liu-Yue has an MBA specializing in investment management and strategy from Georgetown University and a Bachelor of Science in Finance and Marketing from Stern School of Business at NYU. He also completed graduate studies in international management at the University of Oxford, Trinity College.

By Kai Petainen, Forbes,

Have you wondered how one can pick stocks via a quant screen?  Would you like to see the construction of a basic quant screen and the see the top stocks?  Some have heard of the ‘Masked Magician’ in magic, and now I (almost) feel like the ‘Masked Quant’ of quants.  Read onwards …

Pro Tip — For Professionals, this sort of stuff is quite easy to do in FactSet.  Use FactSet, it’s a wonderful tool for quant screening.

In the blog, “The Quant Idea”, I went through a few methods on gathering ideas for ratios and some problems with quant screening.  In the next blog, “The Quant Test”, I went through the basics of testing and creating custom formulas on those ideas.  The next step involves ranking that data and then choosing the stocks via a quant screen at home.  Here are the next steps.

Create the Quant Screen

Using AAII’s Stock Investor Pro, I made a limitation on my universe of stocks by narrowing it down to:

  • Stocks in the USA, but not ADRs
  • Stocks above $5
  • Midcap stocks (those with a market cap between $2 billion and $10 billion)

That gave me a list of about 700 stocks.  With that data, I found (or created) these factors:

  • PE
  • Price / book
  • The Piotroski financial health F-Score
  • The Dechow ‘fudging’ F-Score
  • The sales growth, based on my misinterpretation of ‘GARP
  • The RSST Accrual
  • Insider Buying trades
  • Insider Selling trades

That data was downloaded into Excel and then each factor was given a score from 0 to 10.

Create the Ranking

Long’s Peak. Rocky Mountain National Park

When the data is downloaded into Excel, one needs to decide on each factor and whether or not the high or the low factor is best.  In the case of PE, I will argue that historical evidence shows that lower PEs are best.  As a result, I’ll give the lowest PE stocks a score of 0, and the highest PE stocks will get a score of 10.  I’ll repeat the same idea with the other parameters.  That way, I’ll prefer stocks that have a low PE, low price/book, high Piotroski score, low Dechow f-score, low (my variation) sales growth, low RSST accrual, high insider buying and low insider selling.  At the same time, if the data does not exist, I need to decide on how to grade that data.  For example, with a low PE stock I’ll give a score of 0 and with a high PE stock I’ll give a score of 10; however, if a stock doesn’t have a PE, then I’ll give it a score of 10.  With insider selling, I dislike insider selling, and so higher numbers of insider selling will get a score of 10 and if the data doesn’t exist (there was no insider selling), then I’ll give a score of 0.  How can one do this in Excel?  Here’s an example:


That formula looks at cell CF5 and compares that ratio to all the ratios between CF5 and CF2486.  If the data doesn’t exist, then give the factor a score of 10.  If the data does exist, then rank the formula from a score from 0 to 10.  In this case, higher numbers are better.

Here’s another example, but this is focused on lower numbers


Create an Overall Rank

Sunset over Rocky Mountain National ParkSunset over Rocky Mountain National Park

Next, calculate an overall rank for each stock.  In the example that I gave, each ratio is ranked equally, but there are a number of different ways to ‘weigh’ each parameter.  That is, each ratio can be given a different weighting.  Here are a few strategies:

  • Use equal-weighting on each ratio (as I did in this example)
  • Use equal-weighting on each ratio, but create a style to the method by using different styles of ratios.  In the previous example, it had at least 2 ‘value’ factors and 2 ‘smart money’ factors
  • Weight those ratios that have worked recently, higher than those that worked in the past.
  • Weight those ratios that worked in the past, higher than those that worked recently.
  • Use statistical methods and calculate coefficients for each ratio.
  • Don’t weight anything, but tell your clients that you do. (In my opinion, this is unethical)
  • Weight the ratios that you believe will work in the future, higher.

Using AAII’s Stock Investor Pro, I was able to make the quant screen and download the results into Excel.  If I grab some of the stocks from the top (buy them?) and some of the stocks from the bottom (sell them?), then my result looked like this:

The Final Steps

It’s up to you, or your firm to decide what you do after running a quant screen.  Do you do qualitative analysis?  Do you do more valuation modeling?  Do you look at governance?  Do you include technical analysis?  What about risk?  What about liquidity?  And so on…

I believe the key to quant screening, is that one needs to create a strategy that they believe in.  I hope these blogs have helped you a bit in your pursuit of that strategy…

Based on this quant screen, what stocks are good?

Moonrise over Rocky Mountain National ParkMoonrise over Rocky Mountain National Park

If I take the results from my quant screen, without further analysis and double-checking for data errors… the top ten stocks are:

What stock is at the bottom of the list?

Yes, DNDN.  Recently DNDN had a sizeable drop in share price.  Although there is no guarantee, it’s quite possible that some quant strategies avoided (shorted?) it.

What other stocks appear at the bottom of the list?  (Avoid/short?)

Tags: , , , , , , , , , , , , , , , ,

Post a Comment

Your email is never published nor shared. Required fields are marked *


Subscribe without commenting