Thursday, March 12, 2015

Applying RAND Methods To Finance

The RAND Corporation's CEO spoke at the Commonwealth Club last night and I scored a front row seat.  It helps that the Club's own CEO was once a RAND analyst.  Almost all of RAND's research focuses on cost-benefit planning tools that public policy makers can use.  I noticed that they have published some research on behavioral finance, economics, and energy.  Wall Street can use that stuff free of charge.  Alfidi Capital's never-ending search for financial insights means there is gold to be mined in RAND's methods.

RAND's CEO mentioned the corporation's early work in dynamic programming, linear programming, and game theory.  Analysts in many sectors still use those methods today.  Finance types would have used variations of the Solver program for decades to automate their optimization searches in those methods.  I'm increasingly certain they use Python programs today.  Moving objects around saves time over running long scripting routines.  The operations research that RAND and other government analysts perfected after WWII is now heavily used in Python applications.  Python even handles cost-benefit analysis, RAND's bread and butter.

I was intrigued to hear RAND's optimism over data visualization tools.  Python can do that too, so making data comprehensible to semi-literate policymakers is within every competent analyst's reach.  Technocrats who design our society's future policies now have compelling visual sales tools as the rest of society reverts to pre-literacy levels of comprehension.  I bet RAND also uses the scenario planning method that Royal Dutch Shell and Global Business Network have used for decades.  My web search for jobs involving scenario planning shows a few listings that require proficiency with - you guessed it - the Python language.  There's no escaping coding as literacy these days.  I also noticed a few references to manipulating data stored in Hadoop architectures during scenario analysis.  It's all about Hadoop and Python for finance types who dream of the C-suite for the next decade.

RAND's concerns over threats to the future of objective research are solvable.  Their CEO's main concerns were the short attention spans for digesting longer research products and the political polarization in Washington that discourages objective policy solutions.  First, consider that Arthur McCollum's Eight Action Memo of 1940 summarized US strategic aims in the Pacific with just six pages of text.  Any RAND report with a one-page executive summary is digestible by policy makers who don't code and won't read in depth.  Second, solving the polarization gridlock means presenting research conclusions that are summarized in ways that both liberals and conservatives can understand.  This may require drafting two separate executive summaries for a research product - one written with liberal language, one written with a conservative language - and circulating them separately on Capitol Hill.  Each political audience can then perceive the project summary in language it understands.  Either political interpretation is valid so long as they both point to the same RAND policy recommendations.

I should have applied for the RAND Arroyo Center Army Fellows Program while I was on active duty but it's probably too late in my career to get a slot.  I took a lot of analytical coursework through Army distance learning and my MBA program, and I can still break out my old notes on PERT/CPM and minimax problems if needed.  The stuff comes in handy.  If RAND needs my help, they know where to reach me.  I'm available and affordable.