Type I Errors, Type II Errors, and Slimy Politics
Oct 8, 2008
In yesterday’s presidential debate, Barack Obama attributed the current financial crisis to eight years of Republican misrule and shoddy regulatory practices.
There is truth to this observation; the facts are sitting right there. And yes, we’ve had a Republican administration for the last eight years. But this was a presidential debate – Sen. Obama’s clear implication was that this would not have happened in a Democratic administration.
Let’s go to the numbers. When establishing rules for making binary decisions over and over and over – like approving loan applications – you have to worry about two kinds of errors:
There is a fundamental truth in any process like this: It is almost impossible to minimize BOTH Type I and Type II errors. Tight credit policies minimize default rates, but also are more likely to exclude qualified borrowers. Loose credit policies, like the ones we’re paying the price for now, have the opposite effect.
Now let’s suppose that 20% of the people who were approved for loans because of recent practices actually end up defaulting, an unscientific but frighteningly reasonable estimate. These are the “Type I errors.” But that also means that 80% of those loans are turning out just fine. These are “Type II errors” that were avoided. Put another way, for every family that really shouldn’t have been given a loan, there were four families that did get loans and are now homeowners. [Whether that’s a good tradeoff is not the subject of this piece.]
The real-life implication of minimizing Type I errors is higher lender profit margins and financial stability, because defaults are low. Minimizing Type II errors means more widespread home ownership. Both are worthy goals, but ask yourself: Which party would you associate with trying to minimize each type of error?
For Sen. Obama to attribute the current disaster solely to Republican policies is not just intellectually dishonest, it’s slimy politics.
“Painting with Numbers” is my effort to get people talking about financial statements and other numbers in ways that we can all understand. I welcome your interest and your feedback.
There is truth to this observation; the facts are sitting right there. And yes, we’ve had a Republican administration for the last eight years. But this was a presidential debate – Sen. Obama’s clear implication was that this would not have happened in a Democratic administration.
Let’s go to the numbers. When establishing rules for making binary decisions over and over and over – like approving loan applications – you have to worry about two kinds of errors:
- Type I errors, known as “false positives,” which in this case means approving loans to uncreditworthy borrowers
- Type II errors, or “false negatives,” which here means turning down people who actually are creditworthy
There is a fundamental truth in any process like this: It is almost impossible to minimize BOTH Type I and Type II errors. Tight credit policies minimize default rates, but also are more likely to exclude qualified borrowers. Loose credit policies, like the ones we’re paying the price for now, have the opposite effect.
Now let’s suppose that 20% of the people who were approved for loans because of recent practices actually end up defaulting, an unscientific but frighteningly reasonable estimate. These are the “Type I errors.” But that also means that 80% of those loans are turning out just fine. These are “Type II errors” that were avoided. Put another way, for every family that really shouldn’t have been given a loan, there were four families that did get loans and are now homeowners. [Whether that’s a good tradeoff is not the subject of this piece.]
The real-life implication of minimizing Type I errors is higher lender profit margins and financial stability, because defaults are low. Minimizing Type II errors means more widespread home ownership. Both are worthy goals, but ask yourself: Which party would you associate with trying to minimize each type of error?
For Sen. Obama to attribute the current disaster solely to Republican policies is not just intellectually dishonest, it’s slimy politics.
“Painting with Numbers” is my effort to get people talking about financial statements and other numbers in ways that we can all understand. I welcome your interest and your feedback.
Related Blogs
Other Topics
Other Topics