0:02 Today I'm going to discuss a very
0:04 complicated but crucial topic. The
0:06 financial crisis and how it created
0:09 today's financial system. Its causes are
0:12 multifaceted and complicated and we're
0:14 going to explore it all. Its importance
0:15 can't be overstated. It brought the
0:18 global financial system to its knees.
0:20 The world almost entered a massive
0:23 global depression. By 2006, underwriting
0:25 standards had deteriorated so that
0:27 virtually anyone could get a loan. The
0:29 problem was that everyone involved in
0:31 the process was paid on the basis of
0:34 volume and not loan quality. The more
0:36 volume, the more everyone got paid. So,
0:38 no one had an economic incentive to slow
0:40 the gravy train. And here's where I come
0:56 Hi, this is Steve Eisman and welcome to
0:58 another episode of The Real Eyesman
1:01 Playbook. Today I'm going to discuss a
1:04 very complicated but crucial topic, the
1:06 financial crisis and how it created
1:08 today's financial system. Because of the
1:10 length of this lecture, I'm going to
1:12 break it up into two parts. The topics
1:14 I'm going to explore are the causes of
1:17 the financial crisis, regulatory changes
1:19 that occurred afterwards that have
1:21 permanently changed the industry, why
1:24 did Silicon Valley Bank fail, and was it
1:26 ever a threat to the financial system,
1:29 and finally, the regulatory changes I
1:30 think the Trump administration will be
1:32 making. It's a lot, so let's get
1:35 started. The financial crisis of 2008 is
1:38 now history. Its causes are multifaceted
1:40 and complicated and we're going to
1:43 explore it all. Its importance can't be
1:47 overstated. The crisis was not just a US
1:49 event but global. It brought the global
1:52 financial system to its knees. The world
1:54 almost entered a massive global
1:57 depression. And the actual recession
1:59 that occurred as horrible as it was was
2:02 by far not the worst case scenario. That
2:04 thankfully was avoided. The regulatory
2:06 changes imposed in the financial sector
2:09 afterwards change the financial sector
2:12 and system forever, making it much
2:14 safer, but with unforeseen consequences
2:16 as well. And because of those unforeseen
2:18 consequences, I believe the new Trump
2:20 administration will make changes that we
2:23 will also explore. But first, an
2:26 overview. The financial crisis had four
2:29 interlocking causes. One, too much
2:31 leverage in every financial institution,
2:34 but most importantly in globally
2:36 systemic institutions, which is just a
2:38 fancy term for very large banks and
2:41 investment banks. Number two, a large
2:43 asset class, subprime mortgages that
2:46 blew up. Number three, systemically
2:48 important financial institutions owned a
2:51 lot of that asset class. and four
2:53 derivatives tied the balance sheets of
2:57 large institutions together in a web of
2:59 such complexity that no one knew where
3:02 it started and where it ended. So let's
3:05 start with topic one too much leverage.
3:08 The leverage story is long and I'm going
3:10 to talk about it for a while. To
3:12 understand the leverage story, you need
3:14 to understand that the business model of
3:17 banks is different from other business
3:20 models. Unlike other businesses, banks
3:24 require leverage to generate an adequate
3:27 profit. And a bank's cost of goods sold
3:30 are unknown at point of sale. Now, a
3:33 bank's cost of goods sold are the losses
3:37 that occur from the loans that it makes.
3:39 And when a bank makes loans, it can only
3:42 make an educated guess as to what those
3:45 losses will eventually be. Before I
3:48 explore the too much leverage problem, I
3:50 actually first need to explain how a
3:52 bank works. What does it do? And I'm
3:54 going to spend a long time on this
3:56 because you can't understand the
3:58 financial crisis without understanding
4:00 these crucial fundamentals. The
4:02 traditional bank makes loans.
4:06 Conceptually, I like to say that a bank
4:09 sells you access to its balance sheet
4:12 for a price. And the access is via a
4:15 loan. And the price is the interest rate
4:18 that it charges. I'm going to spend a
4:19 lot of time here because once you get
4:22 this, you are 90% of the way to
4:24 understanding how the financial crisis
4:27 unfolded. And to understand why a bank
4:29 needs leverage to function, let's do a
4:32 thought experiment and start a new bank.
4:36 Let's say bank A, a new bank, raises 1
4:38 billion in equity and the marching
4:40 instructions are for the bank to use
4:44 zero leverage, meaning no deposits. I'm
4:46 going to oversimplify here. So, let's
4:48 say that the bank makes 1 billion in
4:51 loans with the money that it raised, the
4:52 1 billion in equity. Now, that would
4:54 never happen. Every financial
4:56 institution needs some level of
4:59 liquidity or cash for a rainy day. But
5:01 let's just go with my example. Our bank
5:04 A has 1 billion in equity and 1 billion
5:07 in loans. For a bank, the loans it makes
5:10 are its assets. From a mathematical
5:13 perspective, there is no leverage. But
5:16 that means that the leverage ratio is
5:20 one. Here leverage is defined as assets
5:23 divided by equity. And here the equity
5:25 and the assets are the same. Now, how
5:28 profitable can this bank be by making
5:30 loans to good customers at today's
5:33 normal interest rates? Since the bank is
5:35 selling access to its balance sheet for
5:37 a price, the key measurement of
5:41 profitability is return on total assets,
5:45 the ROA, which is net income divided by
5:47 total assets. Here the bank charges
5:50 interest on loans and that is its only
5:53 source of revenue. Then it has operating
5:55 expenses. It has to pay its employees
5:58 and its executives. It has office space,
6:02 computers, etc. It also has to estimate
6:04 future losses on its loans, which is
6:06 called the loan loss provision. And then
6:08 it pays taxes to the government. What's
6:12 left is net income. Take that net income
6:14 and divide it by total assets and you
6:18 get the ROA. So let's assume our bank A
6:23 has an ROA of 1%. Now 1% of 1 billion
6:26 equals $10 million. Trust me in that a
6:30 1% ROA in Bank World is not bad. It's
6:33 not stellar, but it's not bad. JP Morgan
6:36 has more than a 1% ROA, and Cityroup has
6:40 had less than 1% for decades. Bank
6:44 managements are compensated not on the
6:46 ROA of the company, but the return on
6:49 equity, the ROE. And here's the key
6:53 formula, and this is crucial. ROE equals
6:57 ROA time leverage. So here the ROA is
7:00 1%. Since assets and equity are the
7:03 same, the leverage is one times. And so
7:08 the ROE is 1% * 1 equals 1%. So the ROE
7:12 is 1%. Now 1% return on equity is a
7:14 terrible number. So let's try the
7:16 thought experiment again, but with some
7:18 leverage. By the way, we're going to put
7:21 a table up to show each of the examples
7:23 so you can more clearly see the numbers.
7:26 So, in our second example, our bank A
7:28 starts with the same 1 billion in
7:30 equity. And now, let's open a branch and
7:32 bring in depositors. And let's say bank
7:35 A brings in 9 billion of deposits, some
7:38 checking and some savings. The bank is
7:41 now transformed. It has 1 billion in
7:44 equity and 9 billion in deposits. The
7:47 deposits are a form of debt because if a
7:49 depositor shows up and wants his or her
7:52 money, the bank has to give it to them.
7:55 Bank A now has 10 billion of total money
7:57 with which it can make loans. And let's
8:00 say it does make 10 billion in loans.
8:02 First, notice that the leverage ratio
8:05 was now 10 times 10 billion in assets
8:08 divided by 1 billion in equity. Still,
8:10 the only revenue the bank has is the
8:12 interest it charges on this 10 billion
8:15 in loans. But now it has an added
8:18 expense. The interest it pays to its
8:20 depositors. It still has operating
8:22 expenses. It still has to set aside its
8:25 estimate for future losses on its loans.
8:27 And it still has to pay taxes. Let's
8:30 assume that this version of bank A makes
8:36 the same 1% ROA. But here it's 1% * 10
8:39 billion, not 1 billion and that equals
8:44 $100 million. The ROE is now 10%.
8:48 Which is the 1% ROA times leverage of
8:51 10. So from an economic perspective,
8:53 this more levered bank is no more
8:56 profitable than our original non-levered
9:01 bank. They both have an ROA of 1%. But
9:04 because version two is 10 times levered,
9:07 it has an ROE of 10%. Which is not
9:11 terrible versus the earlier terrible 1%
9:15 ROE. Leverage turbocharges the bank's
9:17 profitability. Now let's make the
9:19 example even more extreme. Instead of
9:22 gathering 9 billion in deposits, let's
9:25 take in 99 billion in deposits. Now the
9:28 bank makes 100 billion in loans from the
9:31 99 billion in deposits plus the 1
9:34 billion in equity. The leverage is now a
9:38 100 times. Assuming the same 1% ROA,
9:43 this high octane bank makes 1% ROA times
9:46 100 billion which is a billion dollar.
9:51 The ROE is 100% which is 1% ROA times
9:55 100. So that's how a bank works. As long
9:57 as it's profitable, then the more
10:00 leverage, the higher the ROE. I'm going
10:02 to say that again. As long as a bank is
10:05 profitable, then the more leverage it
10:07 employs, the higher the return on
10:10 equity. There are a lot of lessons to be
10:12 taken from this thought experiment.
10:14 Lesson one, bank executives are
10:17 compensated largely on a combination of
10:20 net income and return on equity. The
10:22 higher the net income and return on
10:25 equity, the more they are paid. Since a
10:28 bank's net income and return on equity
10:30 tends to go higher with more and more
10:34 leverage, the temptation for leverage to
10:36 increase in all banks can be
10:39 overwhelming and often only regulators
10:42 can stop it. Lesson two. In our thought
10:44 experiment, bank A keeps getting more
10:47 profitable as leverage goes up. In all
10:50 three examples, the return on assets was
10:53 1% and the return on equity increased
10:56 solely because of leverage. But banks
10:58 can lose money. How? When they
11:00 underestimate the level of future
11:03 losses. Suppose that because of higher
11:06 losses, our bank loses money and instead
11:10 of generating a 1% ROA, it has a
11:14 negative 1% ROA. To see how too much
11:17 leverage can be dangerous, let's look at
11:21 our highly 100 times levered 100 billion
11:24 bank. -1%
11:28 * 100 billion equals -1 billion. Since
11:30 our bank only started with 1 billion in
11:34 equity, it is now wiped out. So more and
11:36 more leverage is great so long as the
11:39 bank makes money. But if it is too
11:42 levered and loses money, it can be wiped
11:45 out quickly. Lesson three, and I can't
11:48 emphasize this enough, do not take away
11:50 from this thought experiment that all
11:52 leverage in banks is bad. Quite the
11:56 opposite. We want banks to be levered.
11:59 That is their social purpose. Banks take
12:01 deposits that their customers give them
12:04 and recycle that money as loans to other
12:07 customers. We want banks to do that. If
12:09 we did not allow banks to recycle
12:12 deposits or restrain their leverage to
12:14 very low levels, banks would have to
12:17 charge very, very, very high levels of
12:20 interest to generate an adequate return.
12:23 It's because banks employ leverage that
12:26 they can make a decent return on equity
12:29 with only a one-ish return on assets.
12:32 For most businesses, a 1% ROA is awful.
12:36 For a bank, it is adequate so long as it
12:38 can have at least 10 times leverage. We
12:41 want banks to recycle money and
12:43 therefore we want them to have leverage.
12:46 The difficult question is how much
12:48 leverage is too much. We don't want
12:50 banks to have too little and we don't
12:52 want them to have too much. It's kind of
12:55 like little bears porridge. It has to be
12:58 just right. And we explore this later
13:00 on. So, now that I've explained how a
13:03 bank employs leverage and why it needs
13:05 to have leverage to generate an adequate
13:07 return, let's begin to look at the
13:10 crisis. There is no question that by the
13:12 time the crisis began, large banks had
13:15 way too much leverage. How did that
13:20 happen? Between 1997 and 2007, leverage
13:23 in large financial institutions tripled.
13:25 In Europe, for example, the average bank
13:28 leverage ratio climbed from 11 times to
13:32 33 times. In 2007,
13:36 City Bank's leverage ratio was 33 times.
13:39 If you included all the offbalance sheet
13:42 stuff that City Bank had risk for, which
13:44 was not on its balance sheet, the
13:46 leverage ratio was probably over 40
13:49 times. And the same is true for all the
13:51 large investment banks of that era like
13:54 Goldman Sachs and Lehman. And now we
13:56 have to turn to an intellectual concept
13:58 whose importance I cannot overstate
14:02 called risk weighted assets or RWA. This
14:05 one concept might be the most important
14:07 reason for the crisis and its origin is
14:10 completely honorable. In our three bank
14:13 examples, I calculated leverage as
14:15 assets divided by common equity. It was
14:19 just simple math. But imagine two banks,
14:21 each with 10 times leverage and each
14:23 with the same amount of assets and
14:26 equity. Both are levered the same and
14:28 both are the same size, but one bank
14:31 makes only very high risk loans and the
14:34 other only lowrisk loans. How does a
14:37 regulator deal with that? So sometime in
14:40 the late 1980s, the banking system began
14:43 to marry two concepts, leverage and
14:46 risk. And the concept of risk weighted
14:49 assets or RWA was born. Every bank asset
14:52 is given a risk score and that score is
14:55 multiplied by the size of the asset so
14:58 that a bank can calculate not only its
15:01 assets but its risk weighted assets as
15:03 well. So two banks might have the same
15:06 level of assets and equity and both on a
15:09 simple math basis are levered 10 to one.
15:12 But the risky bank might have much
15:14 higher risk weighted assets than its
15:17 assets. And the less risky bank might
15:20 have less RWA than assets. It all
15:22 depends on how much risk each bank is
15:24 taking. So our two banks might both be
15:28 levered 10 to one, but on an RWA basis,
15:30 the risky bank might be levered 20 to1
15:32 and the less risky bank might be levered
15:35 only 5 to one. By the way, up till now I
15:37 have spoken about leverage. But when
15:39 banks discuss this concept, they talk
15:42 about bank capital ratios. This is the
15:45 same thing as leverage. Only the
15:48 numerator and denominator are reversed.
15:51 So leverage is assets divided by equity
15:54 or RWA divided by equity. And bank
15:56 capital ratios are equity divided by
16:00 assets or equity divided by RWA. Just a
16:03 point. Conceptually, the risk weighted
16:06 asset idea makes a lot of sense, but its
16:08 acceptance through banking on a global
16:10 scale had several implications and
16:13 unforeseen consequences.
16:16 Hi, Steve Eisman here. On my weekly rap,
16:18 I try to both teach and convey
16:20 information as objectively as possible.
16:22 I try to make clear what are the facts
16:25 and what are my opinions. But in today's
16:27 media, it's increasingly hard to figure
16:29 out what are the facts and where the
16:32 facts are being shaded by opinion.
16:34 That's why when I look into news events,
16:37 I first go to ground news. Ground News
16:39 is my solution for getting to the facts
16:42 of important stories, but also to see
16:45 how left, right, and center are seeking
16:48 to convey the same exact story. Take for
16:51 example the recent Senate rejection of
16:53 dueling health care bills as the
16:56 Obamacare deadline near. To understand
16:59 the story, I went to groundnews.com and
17:01 clicked on that particular story
17:04 headline. There I immediately saw four
17:08 tabs, left, center, right, and bias
17:10 comparison. When I clicked on the center
17:12 tab, a series of headlines appeared, all
17:15 from centerleaning sources. I could then
17:18 click on any of those headlines and read
17:20 the story at the source. The same
17:22 happened when I clicked on the left tab
17:24 and on the right tab. The bias
17:28 comparison tab showed Ground News's own
17:32 analysis of how all three political
17:35 leanings conveyed the same exact story.
17:37 I find the ground news system enormously
17:40 helpful because it allows me to easily
17:42 separate the facts from opinions. I use
17:45 Groundnews and I recommend you try it
17:52 to get 40% off their unlimited access
17:54 Vantage subscription. That's groundnews.com/real.
18:00 groundnews.com/real.
18:02 And if you don't mind, use this link to
18:04 get the discount so they know I sent
18:07 you. Regulators and bank CEOs look at
18:10 their leverage or capital ratios on an
18:13 RWA basis and not on a simple math
18:16 basis. How is an asset scored to
18:19 generate its RWA? That is mostly
18:21 delegated to the ratings agencies.
18:25 Something rated AAA has a very low RWA
18:27 score, maybe as low as a singledigit
18:30 percentage. Something not rated might
18:32 have a very high RWA score. or something
18:34 rated very low might have a very high
18:38 RWA score. The score is generated by the
18:40 historic level of losses that the asset
18:44 class generates combined with how
18:46 volatile that loss range can be. So an
18:48 asset class with an historically low
18:51 level of losses and a narrow range of
18:54 losses will have a low RWA. Here are
18:58 some unforeseen consequences. First, as
19:00 we've learned, banks have a built-in
19:02 incentive to increase their leverage so
19:04 that their return on equity goes up so
19:07 that the CEO can make more money. So,
19:09 banks have an incentive to load up their
19:12 balance sheets with lowrisisk weighted
19:14 assets. Mathematically, you can have a
19:17 bank with an everinccreasing absolute
19:20 leverage ratio, but because it keeps
19:22 adding on lowrisisk weighted assets, the
19:25 RWA capital ratio might be largely
19:28 unchanged. Number two, the RWA score
19:31 depends on history. What is the historic
19:34 level and range of losses? That level
19:38 and range are not a law of physics.
19:40 Loans are made by human beings and
19:42 institutions. They have underwriting
19:44 standards. And these standards can
19:46 change. They can loosen and they can
19:49 tighten. If for some reason underwriting
19:51 standards loosen continuously over
19:54 several years, losses will gradually
19:57 climb to levels above their historic
20:01 range. I emphasize eventually because
20:03 losses don't happen overnight and this
20:06 can be dangerous. Suppose there is an
20:09 asset class say mortgages that has a
20:12 historic low level of losses and
20:16 therefore a low RWA score. And suppose
20:18 underwriting standards loosen. It takes
20:21 time for losses to eventually show up.
20:24 So banks could be making mortgage loans,
20:26 assuming the old level of losses still
20:29 apply. They could load up their balance
20:31 sheets with these mortgages and their
20:34 leverage on a simple math basis could
20:37 explode while their leverage on an RWA
20:40 basis stays the same. If losses started
20:43 to climb to high levels and the absolute
20:45 leverage ratio was too high, it could
20:48 turn into a disaster. Another one,
20:51 psychology. I can't overstate this
20:53 point. By loading up a balance sheet
20:55 with low risk weighted assets, a bank
20:58 can kind of game the system. It can
21:00 increase its leverage and increase its
21:02 return on equity. And since all bank
21:04 capital and leverage ratios from a
21:06 regulatory perspective are calculated
21:09 via risk- weighted assets, a bank could
21:12 have a higher leverage ratio on a simple
21:15 math basis but a decent leverage ratio
21:17 on a risk weighted basis. And this is in
21:21 fact what happened from 1997 to 2007.
21:23 Because of this increasing leverage on
21:27 an absolute basis, bank roe kept going
21:30 up. Remember when I said that in Europe
21:32 leverage on an absolute basis went up
21:36 from 11 times to 33 times? Guess what?
21:38 Return on equities tripled as well. This
21:41 meant that the return on assets stayed
21:43 the same. Banks were no better run than
21:45 before. Their return on equity or
21:48 profitability improve because of more
21:50 and more leverage and that was it.
21:53 Ironically, leverage during this period
21:55 on a risk weighted asset basis was
21:58 flattish. So bank CEOs developed a sense
22:01 of being godlike because their return on
22:04 equities kept going higher. In fact,
22:06 they really were not any more profitable
22:08 than they had been. It was just they
22:11 used more and more leverage. And you
22:13 don't need to be a genius to increase
22:16 leverage. You just borrow more. And as
22:18 these bank return on equities kept
22:21 climbing, bank CEOs kept getting paid
22:23 more money. And here is the psychology
22:27 part. An entire generation of bank CEOs
22:30 mistook leverage for genius. Suppose you
22:33 went to a bank CEO in 2006 and predicted
22:36 the coming crisis and told the CEO that
22:38 the entire basis of his business model
22:40 was wrong and that the use of more and
22:41 more leverage was going to cause a
22:44 disaster. How do you think that CEO
22:46 would respond? He might not say this,
22:48 but this is what he would have thought.
22:51 I made $50 million last year. I can't be
22:53 wrong. Again, an entire generation of
22:56 bank CEOs mistook leverage for genius.
22:58 Now, before we move on to the second
23:00 cause of the financial crisis, I want to
23:04 pause here to discuss a related issue.
23:06 After the crisis, some commentators
23:08 argued that deregulation of the banking
23:12 industry caused the financial crisis,
23:13 specifically the elimination of
23:16 GlassSteagall. And I think that's just
23:18 wrong. Glass Eagle was a law passed
23:20 during the depression that separated
23:22 banks from investment banks. It died
23:25 during the 1990s. I believe quite
23:28 strongly that even if GlassSteagall had
23:29 been strictly enforced, the great
23:31 financial crisis would still have
23:34 happened. The entire risk weighted asset
23:37 concept was created independently of
23:40 anything related to GlassSteagall. It's
23:42 the leverage that was created because of
23:45 the RWA concept that was the problem.
23:48 and the same subprime loans would have
23:50 been made regardless of glass steagel.
23:53 So let's now discuss that cause number
23:56 two subprime mortgages as a dangerous
23:58 asset class. Subprime mortgages blew up
24:00 the world. But because almost no
24:03 subprime mortgages are made today, few
24:04 people even remember what this asset
24:06 class was about. Now I have a long
24:08 history with the subprime mortgage
24:11 sector. In the 1990s, I was a sellite
24:13 analyst at Oppenheimer where I covered
24:15 the financial services sector. I covered
24:17 everything that was not a bank or an
24:19 insurance company. My coverage was quite
24:21 broad. I followed investment banks,
24:23 asset managers, Fanny May, Freddy Mack,
24:25 credit cards, and several specially
24:28 finance companies. And I also covered
24:30 the subprime mortgage sector. Back then,
24:33 it was a small sector making mostly home
24:35 equity loans to subprime borrowers who
24:38 use the money to pay bills and or to pay
24:40 off other forms of debt. Now, who is a
24:42 subprime borrower? technically a
24:46 borrower with a credit score below 650,
24:48 but that's just a number. The potential
24:50 subprime borrower was the entire lower
24:52 middle class of the United States and
24:54 parts of the middle class as well.
24:57 Starting in the early 1990s, household
24:59 income growth on an inflationadjusted
25:02 basis in the United States stopped
25:03 increasing. It's a problem that
25:06 continues to this day. The reason for it
25:08 are beyond the scope of this podcast.
25:10 Hopefully, we will explore it sometime
25:13 in the future. But statistically, it's a
25:15 fact. The real incomes of most Americans
25:17 stopped growing. So, the only way for
25:19 many Americans to increase their
25:22 spending was to borrow. As a society,
25:24 the lack of middle class income growth
25:27 was and is a serious problem. It's one
25:29 major reason why President Trump is
25:32 president again. And as a country, we've
25:33 never really tackled it. And there was
25:36 no political will to tackle it back in
25:38 the '9s at all. It was far easier to let
25:41 people borrow to increase their spending
25:43 rather than finding ways to improve
25:46 their incomes. Anyway, lending to this
25:49 growing subprime population became a
25:52 gross sector. But prior to the 1990s,
25:55 there was an inherent problem. Companies
25:57 that lent to subprime borrowers were
26:00 small and were looked down upon. They
26:03 had poor rating agency credit ratings
26:05 and so had a hard time getting funding
26:07 to make the loans they knew they could
26:11 make and then securization was invented
26:13 and that changed everything. Now I don't
26:16 want to go into a lot of detail here
26:18 because in the future I will be doing a
26:20 lecture on fixed income where I will
26:22 explore securization in depth. In the
26:25 pre-securization world, the subprime
26:28 lender would raise debt and use those
26:31 proceeds to make loans. The interest on
26:33 its loans was higher than the interest
26:36 on its debt, and it made a spread or net
26:38 interest margin. But because of bad
26:41 ratings, a subprime lender's ability to
26:43 raise debt was limited. In a
26:46 securization, however, a company might
26:49 originate $1 billion in subprime loans,
26:51 package the loans, and put them into a
26:54 securization. The ratings agencies would
26:57 give the securization and not the
26:59 company a credit rating, and the
27:01 interest on the loans in the
27:03 securization would be higher than the
27:06 interest paid on the securizations. So,
27:07 the subprime mortgage company would make
27:10 a spread or a net interest margin
27:13 through the securization. Generally, the
27:15 credit ratings on the securizations
27:17 created a market for these companies to
27:20 raise funding that was far bigger and
27:22 more liquid than their old method of
27:24 raising debt on their balance sheets.
27:27 Suddenly, funding was no longer a
27:29 constraint on growth. And these
27:31 companies began to really grow. And it
27:33 wasn't just subprime mortgages that
27:35 grew. It was also subprime credit cards
27:37 and subprime auto loans as well. Now,
27:39 the first pure subprime mortgage company
27:42 went public in 1992 and others soon
27:44 followed. Some names from back then were
27:47 the Money Store and Ames Financial. In
27:50 1993, the subprime mortgage industry
27:53 generated only 20 billion in loans. By
27:56 1998, the industry was producing about
27:59 150 billion per year. Growth had been
28:02 explosive. But in 1998, the industry
28:04 blew up. It blew up because of bad
28:06 accounting methodology. And it's a long
28:09 story and I'm not going to go into it
28:11 here. Suffice to say that by the end of
28:15 1999, many companies were bankrupt or
28:17 retrenching rapidly and that was the end
28:20 of the subprime mortgage sector 1.0.
28:22 There was a recession in the United
28:25 States in 2000 and when it ended, the
28:28 Fed led by Alan Greenspan had cut the
28:31 Fed funds rate to 1% which back then was
28:34 revolutionary but today seems kind of
28:37 quaint. Back then, this low Fed funds
28:39 rate created a problem for bond
28:42 investors. Bond investors need yield,
28:45 but it was hard to find in a 1% rate
28:48 world, and Wall Street wanted to supply
28:50 it. So, one potential supply would be
28:52 subprime mortgage loans because they
28:54 charged high rates. But the industry
28:57 barely existed anymore. So, Wall Street
28:59 brought new companies public. Companies
29:01 like New Century went public, but not
29:03 all subprime mortgage companies went
29:05 public. Some of the largest like
29:07 Americaest remained private. This
29:09 industry was a gold mine for Wall
29:11 Street. They brought the companies
29:13 public. Then the mortgage companies
29:15 would originate subprime mortgage loans
29:17 and sell those loans in bulk to a Wall
29:20 Street investment bank or to a bank. The
29:22 investment bank or bank would package
29:24 those loans into all different kinds of
29:27 securizations and sell those loans to
29:30 investors throughout the world. The
29:32 lender, the Subprime Mortgage Company,
29:34 made money three to four points it
29:36 charged the consumer and then made
29:39 another two or three points when it sold
29:42 those loans to Wall Street. So five or
29:45 so total points on 1 billion in loans
29:48 equals 50 million in revenue. Wall
29:49 Street would sell those loans via
29:52 securization at a markup as well to
29:55 investors all over the world. Everyone
29:57 made money. It was quite the gravy
30:00 trade. Funny thing was that many of the
30:02 managements of Subprime 2.0 were the
30:05 same people who managed 1.0. That's why
30:07 I began to think that one day this
30:10 industry would blow up again because it
30:12 had blown up in 1998 and the players
30:15 were mostly the same. Normally people
30:19 don't change. So as early as 2003 I was
30:21 looking for signs of problems. What I
30:23 did not imagine was that the subprime
30:25 mortgage sector would get as big as it
30:29 did. In 2002, the industry started to go
30:31 public and growth took off and every
30:36 year industry originations grew. Why?
30:38 Remember when we discussed earlier that
30:40 banks don't know their cost of goods
30:42 sold at point of sale and what that
30:44 means is that cost of goods sold in
30:47 lending is future losses on loans.
30:49 Emphasis here is future. All a lender
30:52 can do is estimate future losses. It
30:54 won't know if that estimate is right for
30:58 a few years. So a lender lends to what
31:01 is called a risk adjusted yield. It
31:03 charges an interest rate to a borrower,
31:05 subtracts its own interest expense, and
31:07 then subtracts the estimate of future
31:10 losses. Again, it subtracts its estimate
31:13 of future losses. That's the risk
31:14 adjusted yield. Think of it as
31:17 profitability before operating expenses.
31:19 Suppose it turns out that delinquencies
31:21 and losses are much lower than
31:23 anticipated. The lender is not really
31:26 happy here because they conclude that
31:28 their lending standards were too tight.
31:30 They could have had looser standards,
31:33 originated far more loans. Sure, they
31:35 might have higher losses, but they would
31:38 still achieve their risk adjusted yield
31:40 and they would be much much much more
31:43 profitable because of the extra volume.
31:44 And that is what happened to the
31:47 subprime sector in the early 2000s. On
31:49 the one hand, the economy was good. On
31:52 the other hand, middle class household
31:54 income was stagnant. But because of low
31:56 rates, the housing sector was doing well
31:59 and home prices kept increasing. People
32:00 tend not to default on their mortgages
32:02 when they are building equity in their
32:05 homes because of rising prices. So from
32:08 2002 through 2006,
32:10 subprime lenders kept loosening their
32:13 underwriting standards. And every year
32:15 delinquencies kept coming in below
32:18 expectations. So the industry loosened
32:22 more. By 2006, growth had exploded.
32:25 Around 600 billion was being generated
32:27 annually in loan volume, which was
32:31 around 20% of the annual mortgage market
32:33 of the entire United States. A once
32:35 cottage industry had become big
32:38 business. But, and here's the big butt,
32:41 underwriting quality had collapsed. I
32:43 remember when I began to do a deep dive
32:46 on the subprime securization market in
32:49 2006 that I discovered something that I
32:52 found astonishing. At least 50% of all
32:55 loans in the securizations
32:58 were originated via a method called low
33:00 dock no dock. And what do I mean by
33:03 that? A subprime borrower has a credit
33:05 score below 650. Credit scores start to
33:09 lose their efficacy below the 650 level.
33:12 So, a subprime loan needs more careful
33:15 underwriting, not less. But because of
33:17 low delinquencies, the industry went in
33:21 the opposite direction, so that by 2006,
33:23 at least 50% of all loans were
33:25 underwritten with very little
33:27 underwriting. The borrower was asked
33:30 what his or her income was, and that
33:32 income was barely verified or not
33:35 verified at all. The borrower was taken
33:38 at his or her word. I like to say that
33:42 by 2006, mortgage underwriting standards
33:44 were so low that if you could breathe,
33:46 you could get a mortgage loan. And I
33:48 don't even think I'm exaggerating there.
33:51 And now I'm going to get on my soap box.
33:53 There were massive social implications
33:55 resulting from the growth explosion in
33:57 subprime mortgage lending. And this was
33:59 because of the structure of the subprime
34:03 mortgage loan. It was typically a teaser
34:05 with a go-to rate adjustment. That's a
34:08 fancy way of saying that the borrower
34:10 got a low fixed rate for two or three
34:13 years and afterwards the rate went up to
34:16 LIBOR plus 600. Essentially, people got
34:20 a 3% rate for 2 to 3 years and after
34:23 that period they got repriced to around
34:27 9% for the next 27 to 28 years. And
34:30 here's the crazy thing. The lender
34:32 underwrote the loan to the teaser rate.
34:34 Meaning that the lender knew at the
34:37 outset that the borrower could only pay
34:39 the 3% rate and that when the rate
34:42 climbed to 9%, the borrower would
34:44 eventually default. Now, was this just a
34:46 scheme to take people's houses? I don't
34:49 think so, because it is expensive to
34:52 repossess someone's home. The design was
34:54 actually different. In a typical
34:57 subprime loan, the originator charges
34:59 three to four points as a fee. That
35:01 compares to one point or less for a
35:04 prime loan. Now, most borrowers and
35:07 subprime land were unsophisticated. They
35:08 really did not understand that their
35:12 loan would be repriced from 3% to 9% in
35:14 2 or 3 years. And when that deadline
35:16 approached, the lender would then
35:19 contact the borrower and remind them.
35:21 the borrower would freak out because
35:23 there was no way they could pay 9%. So,
35:26 the lender offered to refinance under
35:30 the same terms of 3% for 2 to 3 years
35:32 and then the same go-to rate. Of course,
35:34 this was not free. The lender would
35:37 charge the same three to four points for
35:39 refinancing. And the borrower, however,
35:41 would not actually pay out of pocket
35:44 those three to four points. Instead,
35:46 that amount was added to the principal
35:48 amount of the new loan. So borrowers
35:51 were forced to refinance every few years
35:54 and pay the enormous 3 to four points
35:56 for the privilege and rolling those
35:58 points into the principal amount of the
36:00 loan meant that the subprime borrower
36:03 never paid down any principle on their
36:05 mortgage. Subprime borrowers were placed
36:07 on a treadmill from which they could
36:10 never get off. Now socially this was a
36:11 disaster for the middle and lower middle
36:13 classes of the United States. But from
36:16 the mortgage industry's perspective and
36:18 Wall Street's perspective, this was a
36:20 gold mine. Every three years or so, the
36:22 borrower would refinance and the lender
36:25 would make another three to four points.
36:26 And when Wall Street would get to
36:28 repackage the same loans in a new
36:30 securization and they sell those
36:33 securizations to clients all over the
36:36 world at a markup, everybody made money
36:39 and everybody was paid on volume. Yet,
36:42 this gravy train functioned only as long
36:44 as borrowers could refinance. If
36:46 something happened to stop refinancing,
36:49 then all borrowers would get repriced to
36:52 9% and defaults would soar. As I said
36:56 earlier, by 2006, underwriting standards
36:58 had deteriorated so that virtually
37:00 anyone could get a loan. And here's
37:02 where I come into the story. Having seen
37:05 the industry blow up in 1998, I was
37:07 waiting for years for the industry to do
37:09 it again. And that's where securization
37:11 actually was an enormous help to me. The
37:14 subprime mortgage industry securitized
37:17 100% of its loan volume. The securities
37:19 were rated by the ratings agencies and
37:22 every month the agencies put out data on
37:24 each securization.
37:27 Every month. Each securization disclosed
37:30 its 30-day, 60-day, 90-day delinquencies
37:33 for the month as well as repossessions
37:35 and losses per month. This was a
37:37 treasure trove of information. The data
37:39 was sort of public, not public to
37:41 everyone, but if you paid a subscription
37:43 fee to the ratings agencies, which we
37:45 did, you got it. And this data are
37:48 incredibly granular because it comes out
37:51 every month on every securization. For
37:53 analytical purposes, the key was to see
37:55 if there was any deterioration of credit
37:58 quality over time. In other words, my
38:00 partners and I would compare the
38:02 delinquency numbers for multiple
38:06 securization for let's say month 10 of
38:09 each securization. We wanted to see if
38:12 the 2006 securizations had higher
38:14 delinquencies in the same month than
38:17 earlier securizations. And what we found
38:20 was that for the 2006 securizations, the
38:22 early stage delinquencies were far far
38:25 far higher than securizations of any
38:28 prior year. This confirmed what we had
38:31 heard anecdotally that underwriting
38:33 standards had deteriorated terribly and
38:36 by the summer of 2006 it was pretty
38:38 clear that something was wrong. Now if
38:40 the mortgage industry and Wall Street
38:42 had appropriate incentives they would
38:44 have tightened underwriting standards
38:46 immediately and the problem would have
38:49 largely been contained. The problem was
38:51 that everyone involved in the process
38:53 was paid on the basis of volume and not
38:56 loan quality. The more volume, the more
38:58 everyone got paid. So no one had an
39:00 economic incentive to slow the gravy
39:03 train. The industry kept originating and
39:05 securitizing and delinquencies kept
39:08 climbing. The book The Big Short and the
39:11 movie starts in the spring of 2006. Now,
39:13 I'm not going to go through the book,
39:15 but here are a few highlights. We were
39:17 short the equity of a whole bunch of
39:20 public subprime mortgage companies. the
39:22 stocks. We were very convinced that the
39:23 industry was going to suffer enormous
39:27 losses, but the public subprime mortgage
39:29 companies like New Century were not
39:32 large cap stocks and trading was illquid
39:34 and the cost of borrow to short these
39:37 stocks was huge in some quick cases as
39:40 high as 20% annually. We simply could
39:42 not safely allocate sufficient capital
39:44 to shorting these stocks, especially
39:46 given that they were highly shorted. And
39:48 as depicted in the movie, that's when we
39:50 started to learn how to short subprime
39:53 securizations, which was a far bigger
39:55 and more liquid market. How do you short
39:57 a subprime securization? This is
39:59 something I'll talk about when I get to
40:01 the topic of derivatives. So, let's
40:03 pause on that one. Anyway, we began
40:06 shorting subprime securizations in the
40:08 fall of 2006 and kept shorting till
40:12 around July 2007. People sometimes ask
40:13 me how did I have the guts to remain
40:15 short when the hold was telling me that
40:17 everything was fine. Timing is
40:19 everything in life and here my timing
40:22 was kind of perfect. From the fall of
40:25 2006 till the summer of 2007 the
40:28 subprime credit data kept getting mostly
40:30 worse. Every month we would check the
40:32 data and feel good. The one period that
40:34 might have given us pause was in March
40:38 and April of 2007 when the data looked a
40:40 little better. But my partners and I had
40:43 a long history of researching consumer
40:45 lending companies. We knew that in March
40:47 and April of every year, consumers get
40:50 tax refunds and pay down some debt. So,
40:52 of course, the credit data looked
40:54 better. So, the securization market
40:56 experienced a rally, but we just use
40:58 that as an opportunity to put on more
41:01 positions. By the summer of 2007, credit
41:04 data was so bad that no one could be
41:07 remain in denial. investors all over the
41:09 world decided universally that they
41:11 would no longer buy any kind of subprime
41:14 securization, a buyer strike. And that's
41:16 when it all began to really unravel.
41:19 With an investor buyer strike in place,
41:21 Wall Street could no longer sell
41:24 subprime securizations to anyone. And if
41:26 Wall Street could not sell to investors,
41:28 then it had to stop buying loans from
41:30 subprime originators. And if subprime
41:32 origs could not sell their loans and
41:35 they had to stop making them. And if
41:36 subprime originators stopped making
41:38 loans, then consumers could no longer
41:41 refinance. And when consumers could no
41:43 longer refinance, their loans started to
41:46 repric to the go-to rate of around 9%,
41:48 which they could not afford. And then
41:50 the credit picture got even worse. All
41:53 this happened by late summer 2007, which
41:56 meant that the financial crisis was now
41:58 inevitable. Why?
42:01 And on that cliffhanger, I end part one
42:03 of this lecture. If you want to learn
42:06 the rest, wait for next time when I drop
42:08 the second half of this lecture. See you
42:16 This podcast is forformational purposes
42:19 only and does not constitute investment
42:21 advice. The hosts and guests may hold
42:23 positions in stocks discussed. Opinions
42:24 expressed are their own and not
42:26 recommendations. Please do your own due
42:28 diligence and consult a licensed
42:29 financial adviser before making any
