Metaeconomics - The Panic of '08

In my last post I introduced the concept of Metaeconomics.  Microeconomics is the study of the individual.  Macroeconomics is the study of groups.  Metaecomics is the study of the system as a whole.

So why do I call it the "Panic" of '08?  In the 1800's, before the establishment of our modern banking system anchored by the Federal Reserve Bank, there was an event that happened about every ten years.  Out of nowhere, it would seem, a number of people would all of a sudden become concerned about the financial stability of a bank.  They would rush to take their money out "in a panic".  This is called a "run on the bank".  This very action would result in the event the people feared.  The bank run would force a bank out of business and investors and depositors would lose all or a large part of their money.  One bank run would result in one bank failure.  But once one bank failed due to a run people would become concerned that other banks were vulnerable.  This would result in other runs and other failures.  And so on.  Often many backs would be sucked in before things subsided.  And after the bank failed each bank's books would be gone over.  It turns out that in a lot of cases the bank was perfectly sound.  If there had not been a run there would not have been a failure and there was really no reason for the run in the first place.  But people had panicked.  And that became the nickname for these events, "bank panics".

Once a panic had gotten started and had gone on for a while the phenomenon would burn itself out and things would settle down and return more or less to normal.  Of course a lot of people would have been "wiped out".  As I said, a panic that was big enough to involve enough banks to attract widespread notice would take place about every 10 years.  So after a few panics pretty much everyone knew how they worked.  This put people on "hair trigger alert" and they would start a run at the slightest pretext "just to be on the safe side".  Panics were horrible and they did great damage to the economy.  Eventually the "Fed", the Federal Reserve Bank System, was put into place.  And in the Great Depression the Federal Deposit Insurance Corporation was also put into place to insure deposits held in federally chartered banks.  The combination of the Fed and the FDIC was supposed to do away with panics.  You wouldn't be wiped out because your deposits were insured by an arm of the Federal Government.  So there was no longer a reason to panic and participate in a run on the bank.  And for many decades the system worked.  But the events of '08 bear a striking resemblance to a nineteenth century bank panic so that's the term I am going to use.

So that's one digression.  I am now going to go on a second digression and discuss cascading failure.  Then I will get back on track and take a look at the Panic of '08 from a Macroeconomic perspective.  Finally, I will tie it all together by moving to a Metaeconomic perspective and applying what we learned about cascading failure to the events of '07-08.  You've now been warned.

Cascading failure

Let's take a look at our national electric power grid.  We are going to do a very shallow dive so don't panic.  When looking at complicated interconnected systems engineers talk about sources and sinks.  The places where the power comes from are called sources and the places the power goes to are called sinks.  In the case of the "grid" the sources consist of power generating stations, things like hydroelectric dams and power plants.  The power plants can be fueled by coal, natural gas, nuclear power, wind, whatever.  It doesn't matter.  If it produces power it's a source.  Sinks also come in lots of flavors.  They can be small sinks like houses or large sinks like chemical plants, other large manufacturing facilities (airplane manufacturing facilities are popular in my part of the country), anything that uses a lot of power.  Or they can be somewhere in the middle, things like office buildings or apartments.  If it consumes power it's a sink.  Then connecting everything together we have "power distribution" lines.  These are specialized wires that move power from one place to another.  All these parts are connected together by specialized facilities called "distribution yards".  And that's all the technical detail we need to be familiar with.

Now, let's assume that for whatever reason lots of power is being used right now.  That means that the sources are cranked up putting out lots of power and the distribution lines are moving lots of power from one place to another.  Now let's say something goes wrong.  Say a power distribution line gets overloaded and "trips out".  Or let's say something breaks in a distribution yard and it trips out.  Why do I say "trips out"?  Because there are circuit breakers all over the place.  If a circuit breaker detects that too much power is moving through it then the "trip" mechanism kicks in and the breaker "trips out".  The thing that goes wrong can even be a breaker tripping out when it shouldn't.  Anyhow, something goes wrong.  What happens?

Electricity behaves a lot like water.  If it can't go this way it goes that way.  So if a distribution line trips out the electric load instantly switches to any other distribution lines that are still hooked in.  If part of a distribution yard trips out the electric load instantly switches to the other parts of the yard or it finds distribution lines and goes to another yard.  This actually does not happen instantly but it happens pretty quickly.  A second is a long time for electricity.  So now the load has just been increased in other parts of the power distribution system.  What happens?  And the answer is "it depends".

It depends on how much load the components now have to handle.  If the component was lightly loaded before then it will be able to handle the new added load.  So nothing will happen.  No one other than the people who monitor the grid will even know something has happened.  But remember we said lots of power is being used right now.  That means that the affected components are probably running at or near capacity.  The additional load will most likely put them into overload.  The overload will cause circuit breakers to trip out.  The electric load will instantly try to find a new path.  This will most likely cause more components to overload and more circuit breakers to trip out.  This will keep going much like the bank panics used to spread in the 1800's.  Eventually the failures will spread to parts of the grid that are not running at high capacity.  Or grid operators will manually trip breakers to break connections between one part of the grid and others.

We have had a number of these wide spread outages over the years.  They happen in the northeast more often than in other parts of the country because a lot of population and industry is concentrated there.  But, depending on the size, they happen everywhere.  Relatively small ones happen every time we have bad weather, blizzard, hurricane, whatever.  They all start out as a small problem.  Typically one component that is not supposed to trips out.  This causes other components to overload and trip out.  The problem then "cascades" from one component to the next to the next as component after component overloads and trips out.  Hence the name, cascading failure.

To have a cascading failure you need a lot of components connected tightly together and dependent on each other.  A lot of systems have this characteristic.  So how do you avoid cascading failures?  You build in more capacity than you need under normal circumstances.  Bridges and airplanes, for instance, can and sometimes do suffer cascading failures.  Engineers know this.  So they build each component stronger than it needs to be to handle the load it is supposed to carry.  So if a part of a bridge or airplane fails what is supposed to happen is that the other components have enough reserve strength to handle the additional load.  If this happens the plane keeps flying and the bridge stays up.  But sometime the new load is too much or, for whatever reason, the other components are not as strong as they should be (they may have become corroded over time, for instance).  Then we have a cascading failure and the plane crashes or the bridge falls down.  That's cascading failure.  Now let's get back to the Panic of '08.

Macroeconomic analysis of the Panic of '08

Let's begin at the beginning and that's with the Mortgage market.  And let's do a standard macroeconomic analysis.  Above I talked about panics.  They were an all too well understood phenomenon in the 1800's and the early part of the 1900's.  Then we fixed the problem and people started forgetting about them.  I want to now talk about another well understood phenomenon.  But in this case we haven't forgotten about it.  That is the "bubble".  The bible on bubbles was written in 1841 by Charles Mackay.  I alluded to it in an earlier post (http://sigma5.blogspot.com/2014/07/tech-bubble-20.html).  Unlike panics, bubbles, even though they have been with for even longer than panics, have not been forgotten about.  And the mortgage market was a classic bubble.

Bubbles are a phenomena of markets.  So you need a market.  Was the mortgage market a market?  Yes!  Check.  Next, bubbles go through stages.  The first stage is the pre-bubble stage.  This is the period in which the market behaves normally.  We can check off this requirement off too.  In stage two there is the "clean run up".  This is the stage where, for whatever reason, our market goes straight up.  There are no "corrections" where the market goes down (or the corrections are small and short lived so everybody pretends they didn't happen).  This is the "prep" stage.  Again, this stage was present in the mortgage business.  The next stage is the first stage that is part of the true bubble.  In stage three people start telling themselves and others "this market can only go up and it can keep doing it forever".  Check here too.

In stage four people start saying "this has gone on too long - we are overdue for a correction".  In spite of this the market continues to go up.  This is sometimes referred to as "climbing a wall of worry".  Not to put too fine a point on it but definitely check.  Stage five is similar to the start of the panic or the cascading failure.  Something happens and the market (or at least part of it) goes down.  People notice this and say "it's just a blip - stay in the market".  Initially it is not obvious whether this actually is a blip (see stage two) or whether it is more serious.  Often this stage is of very short duration.  Stage six is when people go into panic mode and start dumping in large volumes.  The people who are early to this stage do just fine.  But prices move down at a speed indistinguishable from instantaneous and lots of people take a big bath.  Finally, after a large amount of carnage, things settle down and we go back to the pre-bubble stage.

So the mortgage market was a classic bubble.  Standard macroeconomic analysis would pick that up and, if we stick narrowly to the mortgage market, standard macroeconomic analysis would have done a good job of giving us an accurate picture of what was going on.  But a lot of attention has focused elsewhere.  There was a lot of fraud and bad behavior involved in the mortgage market.  The fraud consisted of putting people in bad (high interest) mortgages when they qualified for good (lower interest) one.  The bad behavior consisted in "no doc" and other gimmicks to get people into mortgages that they couldn't afford.  It is only "bad behavior" because it was all legal.

Both of these activities had an effect.  They made a bad situation worse.  But you could have had a mortgage bubble even if both were absent.  It might have taken longer to form and it might have been smaller.  But neither action affected whether the mortgage market was a bubble or not.  And the bubble nature of the mortgage market was what did the economic damage.  From a policy point of view and from a humanity point of view both activities should be stopped.  But from an economic point of view neither was very important, except to the people directly involved.  If we do a microeconomic analysis of these people they were horribly harmed.

The next link in the chain was Wall Street.  Wall Street bought these mortgages.  In fact the crappier the mortgage the more Wall Street wanted it.  Why?  Because crappy mortgages came with higher interest rates than good mortgages.  And a high interest rate connected to a "safe" investment was catnip Wall Street was drawn to.  They were so strongly drawn to crap mortgages that they told the mortgage market "bring us more crap".  If you are trying to sell an investment and the customer has two to chose from what will the customer do?  In this case both investments are rated AAA (safer than Fort Knox) but one of them pays 6% and the other pays 8 1/2%.  Would you rather be selling the 6% investment or the 8 1/2% investment?  Thought so!  The problem here was that Wall Street convinced itself (at least that's what they say when testifying under oath) and the customers that both investments were AAA.  In reality the 8 1/2% investment was crap and, in a lot of cases, the 6% investment was crap too.

A traditional macroeconomic analysis of the mortgage market in Wall Street would not have shown up any problems.  If the investments has actually been AAA then everything would have been fine.  But they weren't and that's something macroeconomic analysis is incapable of picking up on.  And, given Wall Street's ravenous appetite for crap mortgages, it is not exactly a surprise that the mortgage industry found very creative (sometimes legal and sometimes not) ways of manufacturing crap mortgages in large numbers.  Crap mortgages generated higher "loan origination" fees so mortgage brokers were happy to do what they could to manufacture more.  The "incentive structure" in the business encouraged them and they responded to that encouragement just like all the models said they would.  So for about five years we got a lot of crap mortgages created.

The next link in the chain (link three, if you are keeping track) was also housed on Wall Street.  Why was Wall Street's appetite for crap mortgages so voracious?  Because they had invented something called Collateralized Debt Obligations, CDOs for short.  The whole process is complex.  When you have a bunch of the "smartest guys in the world" (their characterization of themselves) testifying under oath that they didn't really understand CDOs I think we can take it as a given that they were complex.  And part of this complexity was deliberately created to avoid people (regulators, the business press, elected officials, the public) catching on to what was going on.  I have read a number of books that have covered this sort of thing.  I think I have a pretty good idea what the real story was but I am going to just focus on one thing "tranches".

And I am not going into tranches in detail so you can relax.  I am just going to hit a couple of highlights.  Wall Street would buy up a bunch of mortgages (thousands) and dump them all into a single CDO.  So how does that help?  So far it doesn't but that's where tranches come in.  Wall Street would effectively (the actual details are complex and we don't need them) partition the mortgages in the CDO into buckets called tranches.  Why tranches?  I don't know.  I would guess because "tranche" sounds cool and vaguely foreign whereas "bucket" sounds dull and boring.  Anyhow, what the tranche system did was allow Wall Street to sell parts of the CDO.  They put all the high risk stuff into one bucket.  That made all the rest of the stuff low risk.  If the tranche trick was done well almost all of the CDO became AAA.  They would hide the bad stuff away rather than trying to sell it.  They made so much money selling the good stuff that they could afford to take a bath on the bad stuff.

Now if this is sounding like it wouldn't work, you're right.  But what's going on is I am saving you from a bunch of details.  They didn't actually put each mortgage in the CDO into a specific tranche.  What they did was wait.  If a mortgage went bad they retroactively put it in the "bad" tranche.  From an investor point of view, this worked great.  Wall Street could almost guarantee that the stuff in the tranche you bought would never go bad.  And a lot of buyers skipped over the "almost" and took what they heard as a guarantee.  No one took a bath until every single mortgage in the "bad" tranche went bad.  And that was never going to happen.  Why?  Because it never had happened before (or so said Wall Street).  This whole "tranche" business was how Wall Street, with a little help (see below) was able to manufacture all these AAA investments with high interest rates.  The whole system worked great, until it didn't.  Like link two, all this tranche business was invisible to standard macroanalysis so economists missed it.

Link four rests with the ratings agencies.  There are three big organizations and a couple of small ones.  I'm going to ignore the small ones.  Their job is to do the work on behalf of investors of figuring out just how risky an investment is.  Each agency has slightly different criteria and uses slightly different rules.  But generally speaking a low letter in the alphabet is good where a high letter is bad.  So "A" is best.  But they each subdivide "A" into subcategories.  How they do the subdivision also varies.  But "AAA" has become synonymous with "the best of the best".  Each agency has a rating that is or is equivalent to AAA.  Generally speaking any kind of "A" or "B" rating is considered "investment grade".  Higher letters indicate "less than investment grade" and each agency has a grade that is equivalent to a report card grade of "F".  Securities with this grade are generally referred to as "junk".

Not all junk investments fail and some AAA investments fail so the ratings are not foolproof.  But they are supposed to be pretty good.  An AAA rated investment failing is supposed to be like being hit by lightning.  It happens but it is pretty rare.  And there are funds that invest in junk.  Lots of junk eventually pays off.  In the mean time it usually has a very high interest rate associated with it.  Junk funds bet they can keep their failure rate low enough so that the high interest rate more than covers the losses.  That's probably too much information so let me move on.

The ratings agencies rated a lot of investments AAA that were eventually found to be junk.  A lot of investors are required by laws or regulations to only by "investment grade" securities.  So if your security is not investment grade your pool of potential customers shrinks a lot.  So in most cases it is critical that a security get an "investment grade" rating from the agency that rates it.  And, of course, everybody loves an AAA investment with a high interest rate attached.

Now one fly in the ointment is that the company creating the investment picks the agency that ends uprating it.  The agencies know that so a big priority is to keep Wall Street happy.  And the way you do this is by coming up with a lot of AAA ratings.  The only trick is to do this while maintaining a reputation for doing your work properly.  The way to thread this needle is to rely on something called "due diligence".  This is the standard process for performing a rating.  If you rate something AAA and it turns to junk it helps if you can say "I did my due diligence but the security turned to junk so it must have been for reasons beyond my control".  And, frankly for several years the whole thing worked fine.  The AAA stuff paid just like it was supposed to.  In other words, it behaved like AAA securities are supposed to behave.  Then one day all these AAA securities turned to junk pretty much instantaneously.

So what went wrong?  Isn't the "due diligence" process supposed to look in all the nooks and crannies where future bad news may be lurking?  It is.  But these CDOs were a new thing.  There was no history going back decades to consult when looking for ways they could go wrong.  And remember that one of their attributes was that they were wicked complex.  Now a simple response to this would be to say "these are too new and too complex to rate them AAA".  But if a ratings agency did that then Wall Street would just go down the street to another agency.  As long as one agency was willing to look the other way they were all stuck coming up with a rating anyhow, that or go out of business.  None of them were willing to go out of business.

So what did they do?  First, they held their noses and completely ignored the "too new" problem.  As for the "too complex" problem they went along with the Wall Street solution.  Wall Street had come up with a marvelous computer program.  You put a bunch of data in and out popped a single "risk" number.  If the program popped out a number that translated to "low risk" then the agency said "we're good here" and gave it an AAA rating.

Standard macroeconomic analysis might have been able to turn up the "screwed up incentives" problem, the fact that the issuer selected the ratings agency and paid the fee.  But the "no history" problem and the "too complex" problem were of the type that macroeconomics can't detect.  And let me drill down on the ratings program a little more.

The program was developed by a good mathematician.  In theory there was nothing wrong with it if it was used properly.  But Wall Street immediately threw out the "if it was used properly" part.  The details of how the program worked are very complex but we don't need to go into that.  We can just focus on how you used it.  What you did was pour in a bunch of historical data.  This security had this set of attributes and it went bad.  This other security had a different set of attributes and it paid off just fine.  That sort of thing.  Then, after you had loaded up all the historical data you loaded up the details of the investment you wanted to rate.  The program would churn for a while trying to match up the historical stuff with the stuff for the investment you wanted to rate and pop out a number.  If you think about it for a minute you should see how to game the program to get the result you want.

What the program knows about outcomes depends completely on the historical data you pour in.  If you pour in a lot of historical data covering a lot of situations then the program can do a good job.  But what if you pour in a very limited amount of historical data.  And what if all the historical data you pour in represents situations where only good things happen.  In this latter case it is hard to blame the program for getting it wrong.  There is an old saw in the computer business:  "garbage in - garbage out".  It should come as no surprise to learn that Wall Street poured in only the finest sweet smelling garbage.  So the program would look at the security it was rating and decide that it smelled pretty sweet too.  The ratings agencies should have figured out what Wall Street was up to.  But the only one that seems to have figure this out is Wired Magazine.  They did a nice story outing all this and more.  There was exactly zero follow up.  Neither the mainstream media nor the financial media picked up on the Wired revelation and ran with it.  So this shenanigan plays no part in the story most people tell about why things went wrong.

Let's move on to link five.  It is in some ways the most interesting.  There are lots of mortgage brokers.  There are fewer but there are still many players on Wall Street.  There are only three big ratings agencies.  But still, three is more than one.  Link five is all about just one company, AIG.  AIG is an insurance company.  AIG sells (or at least sold) many lines of insurance.  You can get car insurance or boar insurance or life insurance from AIG.  If you are an airline you can (or at least you used to be able to) lease an airplane from AIG.  But all we are concerned about is one of the many subsidiaries of AIG.  And there are lots of insurance companies out there.  And they sell lots of kinds of insurance.  But this one subsidiary of AIG pretty much locked up the market for a specific kind of insurance.  This kind of insurance is called a Credit Default Swap or CDS.  The key word in CDS is the middle one, "default".  The "CDS insurance" subsidiary of AIG sold a policy that paid off if you bought a security and it defaulted (went bad).

Now if we are talking about junk bonds the chance of default is significant.  That's what makes it junk.  But what if we are talking about "safer than Fort Knox" AAA stuff?  Then we are talking maybe one in a thousand, just to be safe.  So let's say I want to buy a CDS on an AAA security.  How much should I have to pay for it.  Well, if I am AIG and I compute the chance is 1 in 1000 then how about I charge 1 500th of the value of the security.  That's 0.2% of the face value of the security.  If the security has a face value of 1 million dollars then the fee would be $2,000.  That sounds pretty cheap.  And remember, the chances of the security actually defaulting (or so everybody thinks) is probably more like one in 10,000 so the expected profit margin for AIG is better than 50%.  It's a perfectly legal low risk "double your money" scheme.

And by marketing aggressively and getting in early AIG locked up most of the market.  Sure, you didn't make that much money on each transaction.  But it was the McDonalds model:  Sell a lot and even if you don't make much on each one you will do fine when you add it all up at the end of the year.  And AIG did.  They made massive profits in this little division for year after year.  Everyone else wanted to get a piece of the action but AIG managed to hold on to it.  And then, of course, everything went south all at once.

These "one in a thousand" shots turned into "one in ten" shots.  All of a sudden charging 0.2% for CDS insurance looked really stupid.  Now AIG was a really big company.  But it had sold a freakishly large number of these CDS deals.  The amount AIG was on the hook for was, in round numbers, a gazillion dollars.  AIG had a lot of money but it didn't have nearly enough money to pay off all the CDS bets that all of a sudden went wrong.  Let's leave AIG aside for a minute and look at the people who bought CDSs.

They did it for a number of reasons.  In some cases they were risk averse.  For a mere $2,000 you could turn a very small chance of losing a million dollars into zero chance of losing a million dollars.  That sounded like a good deal to a lot of careful investors.  In other cases an investment might look risky to a ratings agency.  If you spent $2,000 (chump change on Wall Street) you could say to a ratings agency "say you are right and this is a risky investment.  We just bought CDS insurance.  So if the investment goes bad the insurance company pays off.  So give us our AAA rating".  And the ratings agency did.  Another situation was what is called arbitrage.  Let's say you can buy gold in London for $1,000/oz. (I am using unrealistic but easy to deal with numbers so you can understand what's going on).  Now let's say you can buy gold for $1,020 on the Chicago Mercantile Exchange.  And let's say you have a lot of money.  Well you buy 10,000 oz. in London and sell 10,000 oz. in Chicago.  That should allow you to clear $200,000, risk free.  You can use the gold you bought in London to cover the gold you sold in Chicago.  This process is called arbitrage.  You find two markets for the same thing.  You look for price differences between the two markets.  You buy in the cheap market and match it with a sale for the same amount in the expensive market.  If you can pull it off it is theoretically a license to print money.

If you are going to get in the arbitrage market you want to look for all the ways things can go wrong.  And sometimes you can buy insurance cheap enough so that you can eliminate one or more ways things can go wrong.  So people in the "arb" game often found it appropriate to buy CDS insurance.  And all these people buying CDS insurance for various reasons is why it was a wonderful business for AIG to be in until it wasn't.  It quickly became apparent that AIG couldn't cover its losses.  That meant that some very large and important companies all of sudden stood to lose giant amounts of money.  And in many cases losing that much money would immediately put them out of business.

And here's the point where it make sense to circle back to this whole idea of cascading failure.  The companies that were caught out by the whole AIG CDS fiasco did a lot of business with other companies.  If they went under, and especially if they went under essentially instantaneously and quite unexpectedly, it quickly became apparent that they would take these other "innocent bystander" (at least in many cases) companies down with them.  These, in turn, could take down more, and those more, in a classic cascading failure scenario.  And that's why the Fed found a way to bail AIG out.  At this point it should come as no surprise to learn that a standard macroeconomic analysis of the CDS market or of AIG would not turn the problem up.

Before moving more firmly on to my metaeconomic analysis let me make one more observation about AIG.  AIG is (or at least was) a U.S. company.  Sure it did business around the world but that's a detail.  Except it is not.  Do you know where the CDS division was located?  London.  Why?  The insurance laws in the U.K. are different than they are in the U.S.  AIG would have had all kinds of reserve requirements, capital adequacy requirements, requirements to maintain certain ratios above or below certain numbers, etc.  AIG could not have operated the CDS business the way it wanted to in the U.S.  The U.K. requirements were quite a bit looser and what it was doing was perfectly legal in London.  You can wire vast quantities of money anywhere in the world in less than a second.  So it did not make it any harder for AIG to do business from London than it would have from Hartford (Connecticut Insurance rules) or in Wilmington (Delaware Insurance rules) or any other "business friendly" state.  But London was the friendliest of them all so London was where they did business.  This is a classic example of "regulator shopping" that I have remarked on elsewhere).

Metaeconomic Analysis of the Panic of '08

Traditional analysis said that the mortgage segment was not big enough to take the entire economy down or even Wall Street.  And if it wasn't for the linkages this would have been true.  But it was tightly linked to Wall Street.  So it could and did cascade the failure.  And thorough CDOs Wall Street was able to cascade the failure all across the U.S.  In fact, in an area that I skipped over, CDOs cascaded the failure to encompass the entire world economy.  A contributing factor to Wall Street's ability to sell CDOs was the behavior of the ratings agencies.  Then there was the link represented by AIG and CDS insurance.  The failure immediately took AIG down.  But in taking AIG down the failure, had the Fed not stepped in, would have cascaded out to individuals and companies all across the country and around the world.

And, in another area I have skipped over, by 2008 big banks were heavily invested in the mortgage market.  So the crashing of the mortgage market put these big banks into serious trouble.  Theoretically Joe public was protected by FDIC insurance.  But at the time FDIC insurance only covered the first $100,000.  Few individuals had that kind of money in a checking account.  But many people had that much in "jumbo" CDs.  And many small businesses and every medium or large business had accounts with balances well over $100,000.  If these big banks had gone under then the impact would have been great.  And by this time the FDIC has managed a number of bank problems in such a way that people with high balances were completely covered.  This had translated to an expectation that money in accounts at Federally Insured banks was 100% safe regardless of the balance.  As a result of '08 panic the FDIC temporarily and later permanently raised the insurance limit to $250,000.  But mostly, one way or the other the government bailed out all the big banks.

And there turned out to be more linkages.  Before '08 there were two kinds of banks, "commercial" banks and "investment" banks.  Commercial banks were covered by FDIC Insurance and (theoretically) restricted as to the kinds of business they could engage in.  Investment banks were not insured and could go full "wild west" if they wanted to.  But it turns out that a surprising number of different types did business with the investment banks.  By the time the crisis played out the government was required to find a way to pull investment banks into the FDIC/Fed fold and bail them out.  Any doubt about this was removed when Lehman Brothers, a classic investment bank, was allowed to go under and all Hell broke loose.  Before Lehman an argument was made, and lots of people believed it, that investment banks could and should be allowed to go under if they misbehaved.  This experience hatched the phrase "Systemically Important Institutions".  In plain English it means "too big to fail".  The government can't allow the institution to fail without crashing the economy.

Metaeconomic analysis turns up several threads.  The most obvious one is that the economy is subject to cascading failures.  Most older economic analysis assumed that segments of the market were independent.  And, more than that, they assumed that they were both internally independent and externally independent.  Let me use mortgages as an example.  By internally independent I mean that the performance of one mortgage was not tied to the performance of other mortgages.  But it turned out that forces that operated at the level of the entire U.S. economy propped up mortgages (prices were going up everywhere so you could always exit a mortgage cleanly by selling the house) then they drove down mortgages (high default rates everywhere due to the '08 Panic drove home values down everywhere causing mortgages everywhere to go under water at the same time).  So mortgages were not internally independent.  By externally independent I mean that a downturn in the mortgage market could cause an economic dip but it would not completely crash other parts of the economy.  But we see that the mortgage market turned out to be tightly linked to Wall Street and down stream linkages tied the mortgage market to the entire world economy. This degree of linkage was literally unimaginable before the Panic of '08.

This means that the standard engineering technique of excess capacity need to be applied broadly to many segments of the economy.  For physical structures like airplanes and bridges it means making every component stronger than they need to be to support calculated loads.  How does this translate into economic terms.  In most cases it means reserves.  In the old days a perspective home owner needed to put 20% down before he got a mortgage.  This meant that even if the mortgage went under there was a lot of money available to cover costs like fees and perhaps a dip in the value of the house before the mortgage issuer lost money.  A home owner with a 20% down mortgage has 5-1 leverage.  For every five dollars the mortgage issuer puts out the home owner has put out one.  A lot of Wall Street was running at 50-1 leverage.  This means that if an investment loses 2% of its value the Wall Street firm is in the red.  This gives the Wall Street firm almost no ability to ride through a dip.  Everything must go up and stay up or the firm quickly gets into trouble.  AIG had leverage along the lines of 500-1.  Given that it is no surprise that they got into trouble as soon as the market turned.

The high degree of linkage we now can see means that each of these segments needs significant excess capacity.  The Panic of '08 demonstrates that segment after segment after segment lacks sufficient excess capacity.  It would be nice if we can look at what has happened since '08 and see excess capacity being added broadly.  But mostly we don't.

Another thread I did not get into above is what I call the political problem.  Most regulation is done by the government.  Governments are by their nature political.  That means that they are buffeted by political winds and respond to the political pressures applied.  Sporadically before and definitely during the Great Depression political pressure said "things are out of control - they need to be reigned in".  This resulted in a lot of regulation being put in place.  In general the regulation worked.  Our economy was stable from the '30s through the '60s.  Then the economy suffered some shocks.  There was the "stagflation" of the Nixon era.  Then there were the shocks caused by large increases in the price of Oil and other forms of energy.  This resulted in a political argument being advanced that "the problem is regulation".  The political response was deregulation.  The effort was first applied in the  transportation industry.  Regulations were lifted covering pricing and routes in the airlines and the long haul trucking business.  This was generally viewed as a success.  Airplane tickets got cheaper and the cost of trucking goods around the country dropped.  This deregulation trend has continued ever since, at least to '08.

Other deregulation initiatives were undertaken.  Some seemed to work out ok.  Other didn't.  Savings and Loan institutions were deregulated in the '80s.  This resulted in the "Savings and Loan Crisis" of the late '80s and early '90s.  The government had to step in and bail out S&Ls as over a thousand of them went bankrupt.  One big initiative was Wall Street deregulation.  This was coupled with a cry to "get the regulators off Wall Street's back".  The Glass-Steagall act that limited the kinds of business commercial banks could transact was repealed in the late '90s, for instance.  Utilities regulation and oversight was scaled back too  The combination of Wall Street deregulation and utility market deregulation resulted in the Enron Scandal.  Enron rigged various wholesale electricity markets and engaged on other examples of outrageous behavior.  But this bad behavior was not enough to keep it afloat.  When it went under various accounting bad behavior was exposed along the way.  This resulted in the passage of a law called Sarbanes-Oxley" or Sox, for short.  Many, including myself thought that Sox didn't go far enough.  But Wall Street immediately mounted an extensive lobbying effort to "roll back" Sox.

And then the Panic of '08 happened.  This has hatched, among other things a piece of legislation called Dodd-Frank.  Again, if anything it didn't go far enough.  But efforts are already under way to roll it back.  Another initiative is the Consumer Finance Protection Bureau, an initiative championed by now Senator Elizabeth Warren.  It's job is to protect consumers from predatory practices by lenders.  Its efforts to reign in the most egregious practices by credit card issuers, "payday loan" lenders, and others is being vigorously opposed by the companies that see, quite rightly, that their ability to pull in large profits is being restricted by this bureau.  No one disputes that before CFPB these practices were entirely legal.  The question is not whether they were legal.  The question is "should they continue to be legal going forward".

There is such a thing as a "captured agency".  This is an agency that is supposed to regulate and monitor an industry.  But only that industry much cares what it does.  Through lobbying and other political efforts an industry can capture an agency.  They turn an agency's mission from doing what it should be doing to doing what the industry wants it to do.  Money and finance businesses have waged a long and successful effort to capture the agencies that are supposed to be regulating and monitoring them.

Almost no one pays enough attention to figure this out.  For the most part our print and broadcast media are now owned by large companies.  They want their "news" operations to be financially successful and think this can be done while they simultaneously do not "speak truth to power" if the power is a business power.  Just look at how vigorously the media chases sensationalism by concentrating on fire trucks, sports, and celebrities.  The mantra from the Watergate era was "follow the money".  But if you look, the money in politics is almost never followed in anything but the most superficial way.  As a result, I am pessimistic that turning toward a more pro-regulatory attitude is in the cards.  But a metaeconomic perspective indicates that it is one of the most important steps we can take to avoid a future panic.

A final observation.  The above is pretty pessimistic.  And it is in line with what most observers I trust think.  But there is a ray of hope that is missing from their prognostications.  That is the "sucker factor".  P. T. Barnum is said to have opined at one point that "there's a sucker born every minute".  If he said it then he would have said it in the late 1800's, roughly 150 years ago.  In any case the sentiment has been with us for a long time.  And before the Panic of '08 it was definitely applicable to people who bought CDOs that later turned into crap.  Before '08 these people believed Wall Street sales people who said "we are working on your behalf".  They also believed ratings agencies who said a security was "safer than Fort Knox" (i.e. AAA or equivalent).

The number of people who now naively believe these kinds of representations has diminished substantially.  Wall Street and the ratings agencies have a credibility problem.  They will continue to have a credibility problem as long as the people they victimized are still around.  And I hope that the generations of people that follow are also trained to bring a boatload of skepticism to the table when dealing with Wall Street.  At the moment people are in "trust but verify" mode when they are dealing with Wall Street.  As long as that continues to be the case our chances of avoiding another panic are considerably better than they otherwise would be.  That is grounds for optimism.