First let's think about a mortgage, suppose you buy a house that costs $300,000 and you put $30,000 down. If we think about your house as an investment, you paid 10% of the cost of the investment. Assuming the true value of the house is $300,000, you have $30,000 equity and you owe the bank $270,000. You made an investment of $300,000 with just $30,000 of your own money, this is leverage.
The good side of leverage is when the value goes up just a little, your investment goes up a lot. Suppose the value of your house goes up by 10%, it is now worth $330,000, but your loan from the bank doesn't change, you still only owe $270,000, meaning your equity is $60,000. With a 10% increase in real estate prices your piece of the pie doubled. Of course the downside is when there is a decrease in value, the same magnification happens. A mere 5% drop in the value of the house cuts your equity in half. A 10% drop wipes you out, and more than that puts you underwater. If you needed to move, selling your house wouldn't cover your mortgage. Bad times.
Banks work in a similar way, they make investments, and the more leveraged they are, the higher their upside return will be, but it also comes with downside risk. The appropriate thing to do is try to balance this as much as possible, lots of leverage for the good times, but not so much that the bad times wipe you out. Unfortunately, with our history of bailouts, the banks don't seem to care much about downside risk, they get the benefit in the good times, and the taxpayers get the downside when things to poorly. The result is that they try to get as much leverage as possible without much regard for how dangerous the whole thing is. There are regulations about how much leverage they can have, but they manipulate their figured and hide debt to skirt that as much as possible (She mentioned this in the book, but hasn't yet gone into detail about it).
So, let's say for the sake of argument, you want to have 20% equity in your investments. So out of a $100 investment, $20 is your own money (you owe someone $80). Now, the value of that investment goes down by $1, so you have $19 of a $99 investment. If you want to keep that 20% equity, you can sell $4 of the investment and pay some of that loan back and you will then have $19 out of $95. Really not that much of a big deal.
What if instead the bank only has 2% equity in investments, they have $2 out of the $100. Now if the value drops to $99, they only have $1 out of $99. To get back to their 2% equity they have to sell down to $50, so they need to sell off $49 of the total investment, almost half. This is obviously a much less stable situation.
Furthermore, imagine that there are a ton of banks with similar investments to this last example. The price of their assets drop by 1%, pushing them to sell half of those assets. But if a bunch of banks all sell half their assets at the same time, it will cause the value to drop further. This will force more selling, pushing the price down more and so on. You can imagine this would spiral out of control quite quickly. I wish the 2% equity was just an extreme example to illustrate the point, but it seems that many of the big banks have pushed it this far.
Obviously there was much more going on that made the crisis so huge, but Anat Admati writes that this was a huge cause. As I said, I've only read about 20% of the book so far, but she seems to be arguing that if this was fixed, if banks were forced to keep their leverage much lower then a lot of the other problems would fix themselves. Makes sense to me.