Thinking like a statistician ... to avoid disaster
Photo of Kaiser Fung courtesy of McGraw-Hill
Kaiser Fung suggests that everyone think like a statistician, and he's not just saying that because he happens to be one. Fung, author of the new book Numbers Rule Your World: The Hidden Influence of Probabilities and Statistics on Everything You Do, says you'll be better positioned in life if you pay attention to numbers -- and more importantly, to what the numbers really mean.
I figured you folks would be amenable to this point of view. Here's more from our recent Q&A:
Q. Why should Americans care about economic statistics? What impact do they have on people's lives?
I think the recent downturn is sort of a great example of that ... especially if you’re nearing or about to enter retirement. We kind of can see that this kind of very deep recession will basically make or break your retirement. So if you think about the fact that the housing bubble and the credit bubble was all out there for us to see, in terms of the statistics, if you hadn’t paid attention and you kept a lot of your money in stock-related investments, then you’re in deep trouble. So it really is a very important thing.
Q. What's your take on housing numbers, from home sales to prices to foreclosures?
I’m not an economist, so I’m going to focus more on how a statistician will look at these numbers. One is, it’s very important to distinguish between aggregated averages and local averages. This happens to be a subject I cover in a chapter of my book, which is, should we be looking at one overall average or should we be looking at a lot of averages?
If you look at foreclosure, it’s certainly not hitting the country evenly. Some places, California, Florida and so forth, are much worse hit. So if you just look at national numbers, they’re not going to mean much to you. Same with house prices. … The fact that there’s so much disparity in these numbers ... tells us that we really need to identify what’s the right set of numbers to consider.
The other thing is, we really should pay attention to what’s the incentive for the different players that could be driving these numbers. If you take foreclosures: Do the banks really want to foreclose on people these days? Because every foreclosure is essentially a loan gone bad, and they’re going to have to take losses on their books. Could it be that the foreclosure numbers are severely under-reported? … Thinking behind what the numbers mean is something that someone who has a sense of statistical thinking would be able to do well.
Q. All things considered, then, are these numbers painting an accurate picture of the housing market?
I think the numbers are what they are. It’s the interpretation of the numbers which really matter. If you look at the foreclosure numbers and you realize what the incentives are, you might want to think that the extent of the problem is really hidden. That the numbers give us a false sense of security, even though they’re pretty bad. In reality, it could be worse. ... And that’s why it’s even more important to be able to interpret these numbers and just not take them at the face value.
Q. Think like a statistician, eh?
Yeah, that’s sort of the gist of my book, which is that there are some very basic concepts in statistics that [don't] really require advanced mathematics or anything for everyone to grasp. And if we’re able to think in that way, we’ll actually be much better positioned to interpret the numbers. ...
Daniel Kahneman, who was the founder of behavioral economics, ... also did a lot of research on statistical thinking. What he found is, it's a completely unnatural way of thinking. Our brains are not wired like that. It takes a conscious effort to switch to that mode of thinking. So it’s important to know when to switch.
Q. What should people be paying attention to? That statistics can be squishy?
The main issue I would say has less to do with the squishiness of the numbers themselves but rather ... unfortunate interpretations, missed interpretations of the numbers. And I think the housing bubble is the biggest example of these things having a gigantic impact on our lives.
If you just look at the housing statistics, the build-up of the bubble is very hard ... to miss. At that point in time, whether you’re looking at mortgage debt or housing prices, there is this hockey-stick phenomenon that is all over the place. We have not seen such numbers for a very, very long period of time. But our government unfortunately was staring at the same numbers as a lot of private economists and they came to the conclusion that this is a natural growth and it’s not anything to worry about. And that interpretation has consequences that we are living with and will still live with for many, many years.
At an individual level, if you think about the investors in, say, Bernie Madoff’s fund, that’s a problem too. If they were to pay a lot more attention to the statements they were shown -- and some people did pay attention to them -- they were able to uncover anomalies.
It’s an interesting thing. The reason why the numbers themselves are generally not the problem is it’s usually difficult to fake numbers. The reason for that is that if we think about the economic statistics, they’re all linked to each other in some way. If you try to fake some numbers and you’re changing the relationship to certain statistics, … it’s not hard to uncover that something fishy is going on. It’s much easier to misinterpret the information by either cherry-picking things [or] ... changing the baseline of what you’re talking about.
A common thing that happens is, people will say, ‘Oh, this particular retailer is now doing really well because its same-store sales growth beat expectations by 1 percent.’ That may be in the headline. But if you look into the paragraph itself, it might say its same-store sales growth dropped 5 percent and people expected it to drop 6 percent. … That’s called shifting the baseline. ... An overarching thing we have to be careful about is, are we getting the numbers to fit a story, or are we letting the numbers tell a story?
Q. Lies, damned lies and statistics?
Most of us believe that the numbers themselves don’t lie. It’s the people who are either lying to themselves or misinterpreting the information for others.
Q. Which is why people should become their own interpreters?
For numbers you think will really affect major decisions in your life, you definitely should look at the numbers yourself. And if you were to rely on other people, make sure you understand what the motivations are of the people interpreting the numbers. If you were to … watch CNBC or Bloomberg or any of these channels, most of the people who are talking are actually inside the industry and they may have a reason to either say good things or bad things about whatever number they’re looking at. Without first thinking about what their underlying motivations are, it will be very difficult to interpret the interpretation.
It’s really not just about the numbers. It’s about the people who are using these numbers.
Q. What else should people keep in mind?
There’s really no such thing as true statistics, as in 100 percent accurate. Ultimately, every statistic is based on a sample. So we have to pay attention to how big these samples are, how these samples are selected. But there never will be a true number that can be uncovered.