Wednesday, February 10, 2016

How to Think About Forecasting Events

(Stockholm Street)

Analyzing real estate investments is a forward looking exercise. In my job, I can't focus on the past, I have to forecast the future to figure out what an asset is worth right now. Subsequently, I think a lot about major events, either politically, economically, demographically, etc... and how they will affect the future and in particular how they will affect commercial real estate.

One event you've probably heard of is the stock market has been tanking since the start of the year and this comes after being relatively flat over 2015. In fact, most indices are back where they were two years ago, so it's been a pretty dramatic decline. This is something I've been thinking a lot about and trying to understand how it will affect commercial real estate. I'll talk more about that in a future post, but right now I want to talk about how to think about forecasting the future.

Thinking of Events with the Right Frame of Mind

Whenever I think about current events and how they affect the future, I try to make sure I'm forecasting outcomes with the correct frame of mind. I'd recommend thinking of things in terms of probabilities, i.e. if event x happens tomorrow, I believe there's a y percent chance event z happens in the next year or so. I'd recommend going even further and assigning probabilities to other outcomes as well, so if event z doesn't happen, these other events could happen and they have these specified probabilities of occurring. The fact is, no one knows what will happen in the future with a certainty. Thinking of the future in terms of a potential set of outcomes with probabilities attached to them is the best way to forecast.

Another important facet to forecasting is the idea that tomorrow is very likely to look like today. To illustrate this point, I'll use a common example used in statistics: Weather forecasting. Weather prognosticators have advanced mathematical models, advanced machinery, and many other tools to help them predict what the weather will be tomorrow. You might think their word is gospel, but what if I told you there is a simple trick that will allow you to predict tomorrow's weather more accurately over the long run? You'd probably think I'm crazy, but the fact is that if you just predict tomorrow's weather will mirror today's weather, you'll be right more often than professional weather forecasters and all their mathematical models and weather instruments. It sounds crazy, but it's true. Over the long run, you'll have a higher hit rate by using this method than professionals.

The point is, tomorrow is likely to look like today. Things don't change all that much, especially on a macro level or over short time periods. So how does this impact forecasting? Well, it means if you're trying to forecast the effects of event x over the next year or two, the highest probability is that there will be little to no change resulting from event x. Does this mean forecasting is pointless and we should just assume nothing will change? No, of course not, where's the fun in that? In fact, what if I told you weather prognosticators are aware of the trick above and still use their less accurate mathematical models and instruments? You'd probably wonder why they'd use a more inaccurate method, right?

The reason is, correctly predicting the extreme weather events is actually far more important than just being slightly more accurate. In California, in the summer if you just predict the weather will be 85 degrees and sunny, then you'll be correct extremely often. However, if you're watching the news, you don't really care if the weather is 85 degrees tomorrow. You're already expecting it to be 85 and sunny, so it's not going to impact your plans if it in fact turns out to be 85 degrees and sunny. What you really want to know is if something is going to be abnormal. Will it rain? Will it be 100 degrees? That'll throw a big wrench in your plans. Forecasting those extreme events is what's really important for weather forecasters to get right, which is why they sacrifice some accuracy and attempt to forecast them.

The same is true for our forecasting purposes. It's important to know that things are more likely to remain somewhat static and not undergo extreme change, but there are huge rewards to be had by being prepared for extreme events / changes. Just think of a real estate investor back in 2007. If they correctly anticipated the 2008 crash, they could have liquidated their holdings and gotten out at the top of the market. Then, in 2012 once the market hit bottom, they could have bought up real estate and made a killing as the market rebounded. Of course, the steady Eddie type investor who believes tomorrow will look like today would have held onto his real estate in 2007, taken a hit over the next few years, but as long as they remained steadfast their portfolio would have rebounded and they would have done fine. But not as well as the forecaster that anticipated extreme outcomes. Again, the point is that even though you'll be right more than wrong if you just assume things won't change, there are big rewards for those who can predict extreme events.

No comments:

Post a Comment