Hurricanes, Correlations and Wall Street

This post is based on an interview I had with Institutional Investor that you can read here.

News flash: New York is a coastal town and the entire Wall Street area is only a few feet above sea level!  For anyone not familiar with NYC geography, Manhattan is an island sandwiched between two rivers (Hudson and Harlem), an estuary (known as the “East River”) and NY Harbor, which is open to the Atlantic Ocean.  Click on the map to see in detail what New Amsterdam looked like in 1660 when it was still controlled by the Dutch.

Lower Manhattan in 1660

Note the wall on the right side of the map (North), where Wall Street now stands. Follow Prince Straet south from the wall until first bridge over the canal.  Investor Analytics is located on the corner of what was Begijn Gracht (Beaver Street) and that main canal, now Broad Street.  The NYSE occupies the corner of Prince (also now Broad Street) and Het Cingel (the Wall), just North of us.  All of this land is in the flood zone, as it was when the Dutch ran the place, and they know a thing or two about dealing with water.  For the past 400 years, this island has been an important part of world commerce, and during that time it has been hit by quite a few hurricanes.  According to NYC’s Office of Emergency Management, Lower Manhattan was completely flooded as far North as Canal Street during a hurricane in 1821, and a category 3 hurricane hit the city in 1938.  I count eight different significant storms hitting or affecting NYC from the OEM’s website: 1821, 1893, 1938, Carol, Donna, Agnes, Floyd, and Sandy.  That’s an average of 4 per century – quite a bit more than the “one in a  hundred” we hear about.  Given the reality that Earth is getting warmer, regardless of the cause, and the Northern Atlantic can support larger storms, we should expect even more storms of such strength or worse to hit NYC and other Northeastern coastal cities.  What’s a business located in such a city to do about it?  We at Investor Analytics have a few ideas, based on the very same tools we offer our clients to avoid financial risk.

One of the very basic ideas in financial risk management is “don’t put your eggs in one basket.”  The buzz word is diversification, and the mathematics behind diversification is investing in things that are uncorrelated.  By measuring the correlations of our clients’ investments, Investor Analytics helps them better understand the risks they face.  The more investments they have that are uncorrelated or anti-correlated, the better they can withstand a financial downturn.  We took that advice about uncorrelation and combined it with another important lesson about decentralization when we formed our own disaster recovery plan that we used very successfully during this most recent hurricane.

Lesson 1: Decentralization

The Internet was our model for physical decentralization.  The Internet was designed so that any one server could be off-line but the network would still function.  It was not the “hub and spoke” model that mainframes used.  It was a distributed, de-centralized model that has no “control center.”  In applying that to our own situation, we recognized that we needed an office but that our primary servers didn’t need to be in the same place so long as we had a high-speed connection to it.  Similarly, our backup should be in another separate location with high-speed connectivity.  What about our Disaster Recovery site?  Post 9/11, many firms established DR sites 100+ miles from their primary office (to be independent of electrical problems, etc), complete with workstations, telephones, bathrooms, and even cots.  Rather than follow that model, we decided to turn to our own advice:

Lesson 2: Diversification

We recognized that, for the most part, each employee lives in a different town with different infrastructure.  Some live near the coast, others in the surrounding hills.  The employees’ houses are powered by different utility companies, and some even have generators.  They all use different telco providers for their homes’ telephones and Internet connections.  That’s diversification of infrastructure and therefore diversification of susceptibility to disasters. In a disaster, some of them would have electricity while others would not.  Some would continue to have Internet, others would not.  That diversification means that a core set of employees would continue to be able to work from home.  We wouldn’t know exactly which set of employees that would be until an actual disaster struck, but the likelihood of everyone being off-line is remotely small (but just in case, we have a crack staff in the UK that can run all operations and address client issues).

So our DR location is not “sitting on mothballs” hours away from our office waiting for that day when it’s suddenly turned on.  Rather, our DR sites are used every day by our employees – because they live there – and the secure connections to our servers are tested just about every weekend.  This type of DR planning incorporates another important aspect of modern life: it doesn’t ask the employees to leave home during an emergency to work hundreds of miles away and stay in hotels.  Rather, it asks them to stay at home and be productive from there.  It allows them to take care of their houses, ensure their basements don’t flood, care for a nearby loved one or do whatever else needs their attention.  Employees who are required to travel to a DR site hours away from their homes during such an emergency will tire of that situation quickly.  I’d rather have people who can be productive around the clock because they know their families and property are safe.  This DR plan even allows us to use it during less severe problems like snow storms.  Rather than having employees travel 3+ hours to/from the office in dangerous snow, working from home becomes a viable option.  The employees appreciate it more and the business benefits from having a more productive employee.

While we dry out from Sandy, we have an obligation to think about how we’re going to cope with the next event.  At IA, we’re examining what went well and not-so-well during Sandy to improve our preparedness, and we’ll continue to look at ways of taking advantage of the benefits of diversification.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s