A while ago, I did some forensics on the GSEs (Fannie Mae and Freddie Mac) which I have never gotten around to writing up formally. The long and short of it is that they were never illiquid when they were taken into conservatorship in 2008. They never needed a penny of federal cash to make bond payments, etc. In fact, the taxpayer bailout, when you account for everything, was basically an accounting gimmick.
Well, this got interesting. If Kling is right and Kevin is wrong, does that even matter at this point? There are still a few lonely souls who are claiming that we still have a housing bubble in terms of prices and quantity of supply, but the dominant condition in most cities is one of scarcity caused by a confluence of bad regulatory conditions linked to zoning & mortgage lending.
I'll skip work and meals today and just sit here hitting refresh until Kevin responds to Kling.
I'm sorry, but this essay is way off base. The problem at Freddie and Fannie was that their cost of issuing new debt was above what they could afford. They were going in a deeper and deeper hole. Technically, they had only a $4 billion line of credit from the government, which was not enough to survive. The longer they went on, the deeper the losses would be, and the more that taxpayers would have to pay to keep them from defaulting.
You are expressing great confidence in an analysis that is just plain wrong.
Sure, with the takeover their cost of debt fell, so they weren't in trouble any more. There were two possible outcomes. The no-takeover outcome, which we did not observe, is that their cost of funds exceeds their interest income, and they go bankrupt. With the takeover, their borrowing costs below their interest income. They become "solvent."
1) Was the business model broken or was it a panic? The fact that they didn't need cash once confidence was restored confirms that they weren't structurally insolvent. They weren't in a position where the backstop required any capital injection.
2) That, in an of itself, is pretty impressive, because this was a 100-year flood sort of outcome, and yet they were never structurally insolvent. They were only, at most, struggling with panic-triggered borrowing costs that, it turned out, were based on overly pessimistic expectations, even more pessimistic than the 100-year flood outcomes that ensued.
3) The 100-year flood outcome was, itself, caused by the federal overseers. This comes from my broader work. The credit losses came from defaults that followed deep, temporary losses in low-tier home prices that is directly related to the abandonment of prime mortgage lending for borrowers with average credit scores.
So, yes, you could say that they wouldn't have survived if we just left them to suffer the fates of the panic. We threw them a lifeline. But, at the same time, we tied concrete blocks to their feet. The panic was not a state of nature. It wasn't inevitable. It was the result of changes in their own lending policies which reduced their own collateral, temporarily, by more than 20%.
So, my point is that there really isn't an accounting debate about whether we bailed out the GSEs. If the broader conclusions I bring to the debate are right, government overseers were both executioner and pardoner. And, in fact, it turns out that even with the burden created by the change in lending standards - which was huge - the basic business model could survive without financial support. Even in that case, simply affirming the guarantee on the bonds to stop the panic was enough and the $200 billion was still an accounting fiction.
Curious… how did the RMBS agglomerators factor into this situation, or did they? My suspicion is the big banks and investment houses wanted Fannie and Freddie sidelined while they hogged the market.
Unconstrained hubris has rewards for everyone. The S&L disintermediation/deregulation driven debacle was a semi distant memory. The baby post doc math whizzes hired by the banks and securitization firms were well warned about the dangers of pushing the limits of their new financial modeling tools while their “smarter” new bosses asked them what could possibly go wrong if these new math tools could easily constrain and manage the pool risks? Not much, of course, until the new big bank and securitization firm lords figured out that nascent corporate speech bait dangled on political hooks in front of baby tea party politicians could result in banking deregulation and teeth pulling at Freddie and Fannie. Thus, with a little political pushing and pulling, the profit driven keel of the ship Perfect Storm was laid.
What little remaining concern for housing policy exited the instant these new securitization lords realized a vast quantity of easy money was simply lying around because the very design of these magical new financial modeling tools could tap a gigantic virgin well of previously unavailable risk constrained capital world wide. Oops!
RMBS ended with DB’s failed ACE 2007 HE-4 and 5 securitizations in late Nov.
Most of the US and world banking system’s available lending capacity was tied up in massive unsalable already aggregated and underwritten RMBS pools and there were no buyers for them. The Fed had to pump billions into the system in one weekend just to keep it afloat.
Imagine that as massive bank trucks full of hamburgers backing to unload at every MacDonald hamburger shop. But there were no hamburger buyers.
Was RMBS a policy decision? Keeping the system afloat was. Who was at fault for the RMBS situation? DB paid a $7.2 B settlement in 2016. Reg Z had a minimal impact on RMBS lending.
That makes sense. You have increased my estimation that those markets would have shriveled away in any case.
As I noted, I see a small rise of about 8% in home prices from 2002 to 2006 which then reversed, which is associated with credit access in general, including the subprime RMBS that had disappeared by mid-2007.
Then, there is a completely separate credit event, which is associated with a completely different set of signals (a change in credit scores on approved mortgages after 2007), which had a measurable effect on home prices (a decline in home prices in every city after 2007 that was correlated with ZIP code income). The policy choices I associate with the post-crisis housing depression are related to the change in prime mortgage lending at the GSEs.
That's one of my biggest frustrations with the conventional wisdom on the crisis. The change in mortgage access associated with the federal agencies in 2008 was much more important than the rise and fall of the RMBS market, and it is generally both conflated with the rise and fall of RMBS and treated as if it is unimportant. Changes in lending at the federal agencies after RMBS had died is associated with a 20% decline in average home values - the entire $5 trillion net loss of wealth - and was highly regressive. And it is unacknowledged in the academic literature.
Cause and effect. The supply of loan capital was essentially infinite relative to “normal” demand from 2002-2006, with additive borrower pools as the bank/Wall Street firms lowered their standards from prime to warm body status. Everyone wanted to buy, but not everyone wanted to sell. Maybe only 1% of all houses traded, maybe a little more (Zillow knows but will they tell you?). My very subjective guess is there were at least 5 buyers for every available home and high demand homes could have 20. Overbids and no asking price offerings became normal. Prices were bubbled up even more from a year earlier, each year.
Loan amounts went up with increased prices, and some were at 95% of the strike price. Fact: The RMBS aggregators never assumed home prices would or could go down, by the way. The aggregators were competing with each other for the same loans; better terms.
Simultaneously borrowers were refinancing to pull equity out, a lot of it. Again, aggregator competition.
The standard assumption by borrowers, aggregators, and sideline watchers was that the pipeline of lender and loan availability was infinite, in both RMBS and CMBS.
Refi was easy, just find another lender. But the pool level profit spreads were thinning down. DB’s last to pools were only 37 or 40 bps, if they closed. No room for errors. The RTC RMBS pools by comparison were about 450-650 bps.
Except, at the lower quality pool since mid 2006 things were not so rosy. The top AAA A1 tranche investors got paid but 10 or 20 levels down didn’t. At first the aggregators could trade out a tranche position with a new buyer, but as the word started getting around there were fewer alternatives. MIT Technology Review published an article Is Subprime the Canary in the Coal Mine in mid 2007 (I think), and a couple of later articles telling a scary story.
By the end of 2007 there were no buyers. For any pool. Implosion!
The post docs knew what was going on but their bosses didn’t. Jamie Dimon was one of the biggest refuseniks.
Values dropped precipitously but not because of lending standards. The pool of warm bodies was 5% of what it was. Sellers needing to sell had few buyers, and those needing loans had few options. But Fannie and Freddie were not the cause of the decline in values or lenders. It was that there were no buyers with easy money. The bubble created by easy money popped and more “normal” market conditions ensued.
Fannie and Freddie were merely servicing the market that existed prior to the bubble, as redesigned by Dodd-Frank, etc., to fit their curmudgeonly punitive views about proper lender and borrower behavior. Dimon should have gone to jail along with 50 other bankers, the heads Fitch, S&P, AIG, etc.
I don’t see Fannie and Freddie programs as anything more than political retribution to earn voter brownie points. It does not qualify as housing policy per se.
Hm. Your descriptions here seem reasonable. I'm surprised to see your closing comments though.
I agree that the RMBS were increasingly for refis. But, I think one thing you have to contend with on your description is that sales peaked in late 2005. For nearly 2 years after that, the strong lending market you describe was associated with a deep decline in both new and existing home sales and a rise in both new and existing inventory, so much of the subprime period was not associated with multiple offers, etc.
And on the Fannie and Freddie part, it's just an empirical fact that the borrowers and homes they services were a relatively stable pool from the 1990s to 2007, and then in 2008 they sharply changed their standards and the number of mortgages made to borrowers with average or below average credits scores declined sharply (this was entirely unrelated to the rise and fall of subprime)
One other thing I would note on RMBS, and I'll be interested to see if this matches with your experience. In BFTGU, I wrote about the S&P subprime downgrades in mid-2007 when those markets were seizing up. I think, even in that case, it was largely a panic. The AAA tranches were trading at deep discounts, and the actual cash returns on them were not that bad, in the end. The initial downgrades were in response to some initial increases in defaults, but mostly they were related to expectations - that unprecedented future declines in home prices would lead to very high default rates later in 2007, 2008, etc.
In a way, you could say that that panic was related to monetary policy because the Fed would have been stimulating by then to encourage a recovery in housing starts if they hadn't been taken in by fatalism about housing. In other words, S&P expected home prices to decline by 8% over the following year, triggering a bunch of negative equity defaults, because everyone correctly inferred that Fed policy, by then, would be aimed at facilitating that sort of decline.
Much like the later experience of the GSEs, there was a panic in the valuations, but when all was said and done, cash flows were better than the panic prices had presumed.
So, to the extent that the country was unanimously engaged in a fatalism about declining home prices, and that fatalism informed Fed policy, it was a policy decision to create a set of expectations that would lead to the S&P downgrades. The defaults were expected before they happened, and there was no motivation at macro levels to stop them.
But, that's all a separate and less direct story than the later GSE story. Subprime could have died, and if Fed policy and federal prime lending in 2008 had been aimed at stability, the $5 trillion loss of wealth wouldn't have happened. The later 2008 federal stuff is much more important in my estimation than the pre-2008 subprime stuff.
Yeah, I agree with that. But, I’m not so sure that RMBS as implemented prior and up to the late Nov 2008 DB offering failure (the last straw) represented any lending policy, except for Greenspan’s much earlier comment about everyone deserves a house and by implication, a loan. Even that comment wasn’t policy, but it did encourage more aggressive pooling in lower quality loans.
The main policy choice was probably the 2007 update to Regulation Z, which may have hastened the end of the subprime and Alt-A markets. I don't have a strong opinion on that. Maybe they would have shriveled away in any case. But, then, what really knocked the wind out of the housing market was that great pressure was put on the GSEs to tighten lending after subprime had disappeared. Once they were taken over, the government had complete control over their underwriting. And various updates to underwriting regulations in the CFPB, Dodd-Frank, etc., permanently put an end to any sizeable market in private mortgage lending with any significant amount of default risk.
The short version is just that by 2008, the federal government had complete operational control of mortgage lending standards and enforced standards that permanently removed 1/3 of former borrowers from the market.
I think the most powerful source of inference in the policy shift that I have is the charts of credit scores, loan size, and home values on new mortgages at Fannie Mae.
The Lucas critique is just that if something is a policy goal, then it ceases to be empirical evidence for anything else.
For example, there may be a lot of things that used to correlate with inflaiton. If those things don't have a correlation with inflation any more, it's just because the Fed policy framework removed variance from inflation. That thing may still be inflationary, all else equal, but Fed policy is designed to change other variables in order to counter the inflation.
Kevin, you and are on the same page about a need for a sound national housing policy. Our perspectives come from opposite sides of looking at the same facts, yours as an academician and mine as a hands on operator with 55 years of incredibly diverse experience in all real estate asset classes, types and locations. I’m working very diligently to walk in your shoes and understand your work.
But, I’m struggling a bit with inferring housing policy from tool usage (like RMBS, or Fannie and Freddie loan portfolios and loan policies, etc.). Hoping you can help me here.
Is there a simple cause and effect description for the housing policies that you see? How and what would change that paradigm and result in a greater supply and more “affordable” housing?
Maybe? If you are referring to the current standards, no. That’s pretty clear, and they are an overreaction to something borrowers didn’t create. It’s earlier times, I think. Like the ones used by RMBS pools.
If I understand what you're asking, I would say that there was a bit of a lending bubble from 2003-2007. It was probably associated with temporary inflation in aggregate American housing values of up to about 8%. That's pretty sizeable. I wish that 8% boost had been the most important factor in American home values.
I have problems with other analysis. Home prices nearly doubled from the mid-1990s. The credit portion of the boom was insignificant. That is obvious now, since the arrest of those markets did not stop further price appreciation. Conventional analysis usually controls away 80% of factors (including differences between metro areas), then sort of pretends that the few % of a price boost they attribute to credit markets can be inferred to be much larger. They are wrong. Supply constraints created 80% of the rise in prices, unrelated to credit markets, and there was a credit boom on top of that which would have been sizeable in any other context, but is insignificant compared to our other problems.
Also, the rise in homeownership happened almost entirely before the subprime boom and the peak of the price boom. The private securitization originations that had extremely high default rates happened when prices had leveled off and home sales were in free fall. The timing doesn't allow it to be an important causal factor in more than a small portion of price appreciation. And, finally, the CDO boom was entirely after all housing trends had turned south, and has been systematically mischaracterized as an example of leveraged risk-taking when it was really the result of a market place desperately seeking low risk destinations for savings in a market where home equity was increasingly seen as a risky investment.
Is it possible that the “poor housing policies” you are focused on resulted from an ingrained institutional default need to fill any policy vacuum with something, or anything, when the then politicians weren’t interested in or even capable of providing any well thought out policies and goals?
My recollection is that the nascent GOP tea party politicians were solely focused on deregulation and the destruction of extant policies while simultaneously loosing the dogs of unfettered free market capitalism as the best means to solve all problems.
Additionally they benchmarked to crisis and poverty, so they only supported unfettered free markets where that would lead to liquidation and unsustainably low market prices.
In the end, what that meant is that their only principle was crisis and poverty. And, it was easy enough for them to tell themselves that where markets wouldn't have led to liquidation and panic, it was because of the federal reserve, Fannie & Freddie, etc.
The Federal Reserve was the wild card that could be used to assert that any non-crisis outcome was the result of non-market interventions. This is still common today.
I’m thinking that “benchmarking to crisis and poverty leading to liquidation and unsustainable low prices” aspect is baked into unfettered capitalism by default. I see it as a consequence not a goal. Ascribing such a dark premise as an intentional act in public policy is unimaginable. However, not minimally footnoting it as a possible outcome is imaginable.
Part of my thinking stems from the unintentional but very real black swan risks that live in long tails. Those risk become increasingly likely as systemic underwriting pressures are increased and unintended consequences from loosely related exogenous variables arise. (Say in a heavily reinsured RMBS pool of lower quality loans, coupled with an increasingly skittish investor market that results in the disappearance of some or most of each of the individually rated tranche’s resale liquidity. The inherent liquidity assumption in each and all pools and tranches is that a relatively inexhaustible supply of new or other pools and investors is always available, which of course was and is a fallacy.
Notwithstanding that liquidityn the surface a tranche or pool liquidity failure could be ascribed to your observation. Looking deeper shows that it isn’t; the failure was because of underwriting aggressiveness and market forces. Neither are policy decisions, unless one can conclude that the mere existence of the securitization tools and underwriting rules are the policies. If so, an effective policy determination would be made with each individual pool much like a theoretical adhocracy would work.
If there are 10 types of securities, ranging from least to most risky, and public policy becomes benchmarked to some sort of failure, then the most risky securities will fail. But, you can't then conclude that their riskiness was the cause of the failure. The policy was the cause of the failure.
There were certainly ill-advised securities. And, they were what failed. I'm not sure if their failure was inevitable. And, in fact, I would argue that the crisis and the recession were so deep because the amount of economic contraction required to make them fail was much greater than what anyone assumed. The recession was Great because there actually weren't a lot of systemically important securities ready to fail at the drop of a hat.
Well, this got interesting. If Kling is right and Kevin is wrong, does that even matter at this point? There are still a few lonely souls who are claiming that we still have a housing bubble in terms of prices and quantity of supply, but the dominant condition in most cities is one of scarcity caused by a confluence of bad regulatory conditions linked to zoning & mortgage lending.
I'll skip work and meals today and just sit here hitting refresh until Kevin responds to Kling.
I'm sorry, but this essay is way off base. The problem at Freddie and Fannie was that their cost of issuing new debt was above what they could afford. They were going in a deeper and deeper hole. Technically, they had only a $4 billion line of credit from the government, which was not enough to survive. The longer they went on, the deeper the losses would be, and the more that taxpayers would have to pay to keep them from defaulting.
You are expressing great confidence in an analysis that is just plain wrong.
If that is true they would have needed to use some of the federal cash injection to pay off bonds. But they didn’t, as you can see in my chart.
Sure, with the takeover their cost of debt fell, so they weren't in trouble any more. There were two possible outcomes. The no-takeover outcome, which we did not observe, is that their cost of funds exceeds their interest income, and they go bankrupt. With the takeover, their borrowing costs below their interest income. They become "solvent."
Yes. There are several layers here.
1) Was the business model broken or was it a panic? The fact that they didn't need cash once confidence was restored confirms that they weren't structurally insolvent. They weren't in a position where the backstop required any capital injection.
2) That, in an of itself, is pretty impressive, because this was a 100-year flood sort of outcome, and yet they were never structurally insolvent. They were only, at most, struggling with panic-triggered borrowing costs that, it turned out, were based on overly pessimistic expectations, even more pessimistic than the 100-year flood outcomes that ensued.
3) The 100-year flood outcome was, itself, caused by the federal overseers. This comes from my broader work. The credit losses came from defaults that followed deep, temporary losses in low-tier home prices that is directly related to the abandonment of prime mortgage lending for borrowers with average credit scores.
https://fred.stlouisfed.org/graph/?g=1ojwW
So, yes, you could say that they wouldn't have survived if we just left them to suffer the fates of the panic. We threw them a lifeline. But, at the same time, we tied concrete blocks to their feet. The panic was not a state of nature. It wasn't inevitable. It was the result of changes in their own lending policies which reduced their own collateral, temporarily, by more than 20%.
So, my point is that there really isn't an accounting debate about whether we bailed out the GSEs. If the broader conclusions I bring to the debate are right, government overseers were both executioner and pardoner. And, in fact, it turns out that even with the burden created by the change in lending standards - which was huge - the basic business model could survive without financial support. Even in that case, simply affirming the guarantee on the bonds to stop the panic was enough and the $200 billion was still an accounting fiction.
Curious… how did the RMBS agglomerators factor into this situation, or did they? My suspicion is the big banks and investment houses wanted Fannie and Freddie sidelined while they hogged the market.
The 2000s scandals do seem motivated, which does make one wonder the source of the motivation.
The way things worked out, it seems to have backfired in every way, doesn’t it?
Unconstrained hubris has rewards for everyone. The S&L disintermediation/deregulation driven debacle was a semi distant memory. The baby post doc math whizzes hired by the banks and securitization firms were well warned about the dangers of pushing the limits of their new financial modeling tools while their “smarter” new bosses asked them what could possibly go wrong if these new math tools could easily constrain and manage the pool risks? Not much, of course, until the new big bank and securitization firm lords figured out that nascent corporate speech bait dangled on political hooks in front of baby tea party politicians could result in banking deregulation and teeth pulling at Freddie and Fannie. Thus, with a little political pushing and pulling, the profit driven keel of the ship Perfect Storm was laid.
What little remaining concern for housing policy exited the instant these new securitization lords realized a vast quantity of easy money was simply lying around because the very design of these magical new financial modeling tools could tap a gigantic virgin well of previously unavailable risk constrained capital world wide. Oops!
RMBS ended with DB’s failed ACE 2007 HE-4 and 5 securitizations in late Nov.
Most of the US and world banking system’s available lending capacity was tied up in massive unsalable already aggregated and underwritten RMBS pools and there were no buyers for them. The Fed had to pump billions into the system in one weekend just to keep it afloat.
Imagine that as massive bank trucks full of hamburgers backing to unload at every MacDonald hamburger shop. But there were no hamburger buyers.
Was RMBS a policy decision? Keeping the system afloat was. Who was at fault for the RMBS situation? DB paid a $7.2 B settlement in 2016. Reg Z had a minimal impact on RMBS lending.
Thanks for this.
That makes sense. You have increased my estimation that those markets would have shriveled away in any case.
As I noted, I see a small rise of about 8% in home prices from 2002 to 2006 which then reversed, which is associated with credit access in general, including the subprime RMBS that had disappeared by mid-2007.
Then, there is a completely separate credit event, which is associated with a completely different set of signals (a change in credit scores on approved mortgages after 2007), which had a measurable effect on home prices (a decline in home prices in every city after 2007 that was correlated with ZIP code income). The policy choices I associate with the post-crisis housing depression are related to the change in prime mortgage lending at the GSEs.
That's one of my biggest frustrations with the conventional wisdom on the crisis. The change in mortgage access associated with the federal agencies in 2008 was much more important than the rise and fall of the RMBS market, and it is generally both conflated with the rise and fall of RMBS and treated as if it is unimportant. Changes in lending at the federal agencies after RMBS had died is associated with a 20% decline in average home values - the entire $5 trillion net loss of wealth - and was highly regressive. And it is unacknowledged in the academic literature.
Cause and effect. The supply of loan capital was essentially infinite relative to “normal” demand from 2002-2006, with additive borrower pools as the bank/Wall Street firms lowered their standards from prime to warm body status. Everyone wanted to buy, but not everyone wanted to sell. Maybe only 1% of all houses traded, maybe a little more (Zillow knows but will they tell you?). My very subjective guess is there were at least 5 buyers for every available home and high demand homes could have 20. Overbids and no asking price offerings became normal. Prices were bubbled up even more from a year earlier, each year.
Loan amounts went up with increased prices, and some were at 95% of the strike price. Fact: The RMBS aggregators never assumed home prices would or could go down, by the way. The aggregators were competing with each other for the same loans; better terms.
Simultaneously borrowers were refinancing to pull equity out, a lot of it. Again, aggregator competition.
The standard assumption by borrowers, aggregators, and sideline watchers was that the pipeline of lender and loan availability was infinite, in both RMBS and CMBS.
Refi was easy, just find another lender. But the pool level profit spreads were thinning down. DB’s last to pools were only 37 or 40 bps, if they closed. No room for errors. The RTC RMBS pools by comparison were about 450-650 bps.
Except, at the lower quality pool since mid 2006 things were not so rosy. The top AAA A1 tranche investors got paid but 10 or 20 levels down didn’t. At first the aggregators could trade out a tranche position with a new buyer, but as the word started getting around there were fewer alternatives. MIT Technology Review published an article Is Subprime the Canary in the Coal Mine in mid 2007 (I think), and a couple of later articles telling a scary story.
By the end of 2007 there were no buyers. For any pool. Implosion!
The post docs knew what was going on but their bosses didn’t. Jamie Dimon was one of the biggest refuseniks.
Values dropped precipitously but not because of lending standards. The pool of warm bodies was 5% of what it was. Sellers needing to sell had few buyers, and those needing loans had few options. But Fannie and Freddie were not the cause of the decline in values or lenders. It was that there were no buyers with easy money. The bubble created by easy money popped and more “normal” market conditions ensued.
Fannie and Freddie were merely servicing the market that existed prior to the bubble, as redesigned by Dodd-Frank, etc., to fit their curmudgeonly punitive views about proper lender and borrower behavior. Dimon should have gone to jail along with 50 other bankers, the heads Fitch, S&P, AIG, etc.
I don’t see Fannie and Freddie programs as anything more than political retribution to earn voter brownie points. It does not qualify as housing policy per se.
Hm. Your descriptions here seem reasonable. I'm surprised to see your closing comments though.
I agree that the RMBS were increasingly for refis. But, I think one thing you have to contend with on your description is that sales peaked in late 2005. For nearly 2 years after that, the strong lending market you describe was associated with a deep decline in both new and existing home sales and a rise in both new and existing inventory, so much of the subprime period was not associated with multiple offers, etc.
And on the Fannie and Freddie part, it's just an empirical fact that the borrowers and homes they services were a relatively stable pool from the 1990s to 2007, and then in 2008 they sharply changed their standards and the number of mortgages made to borrowers with average or below average credits scores declined sharply (this was entirely unrelated to the rise and fall of subprime)
One other thing I would note on RMBS, and I'll be interested to see if this matches with your experience. In BFTGU, I wrote about the S&P subprime downgrades in mid-2007 when those markets were seizing up. I think, even in that case, it was largely a panic. The AAA tranches were trading at deep discounts, and the actual cash returns on them were not that bad, in the end. The initial downgrades were in response to some initial increases in defaults, but mostly they were related to expectations - that unprecedented future declines in home prices would lead to very high default rates later in 2007, 2008, etc.
In a way, you could say that that panic was related to monetary policy because the Fed would have been stimulating by then to encourage a recovery in housing starts if they hadn't been taken in by fatalism about housing. In other words, S&P expected home prices to decline by 8% over the following year, triggering a bunch of negative equity defaults, because everyone correctly inferred that Fed policy, by then, would be aimed at facilitating that sort of decline.
Much like the later experience of the GSEs, there was a panic in the valuations, but when all was said and done, cash flows were better than the panic prices had presumed.
So, to the extent that the country was unanimously engaged in a fatalism about declining home prices, and that fatalism informed Fed policy, it was a policy decision to create a set of expectations that would lead to the S&P downgrades. The defaults were expected before they happened, and there was no motivation at macro levels to stop them.
But, that's all a separate and less direct story than the later GSE story. Subprime could have died, and if Fed policy and federal prime lending in 2008 had been aimed at stability, the $5 trillion loss of wealth wouldn't have happened. The later 2008 federal stuff is much more important in my estimation than the pre-2008 subprime stuff.
Yeah, I agree with that. But, I’m not so sure that RMBS as implemented prior and up to the late Nov 2008 DB offering failure (the last straw) represented any lending policy, except for Greenspan’s much earlier comment about everyone deserves a house and by implication, a loan. Even that comment wasn’t policy, but it did encourage more aggressive pooling in lower quality loans.
The main policy choice was probably the 2007 update to Regulation Z, which may have hastened the end of the subprime and Alt-A markets. I don't have a strong opinion on that. Maybe they would have shriveled away in any case. But, then, what really knocked the wind out of the housing market was that great pressure was put on the GSEs to tighten lending after subprime had disappeared. Once they were taken over, the government had complete control over their underwriting. And various updates to underwriting regulations in the CFPB, Dodd-Frank, etc., permanently put an end to any sizeable market in private mortgage lending with any significant amount of default risk.
The short version is just that by 2008, the federal government had complete operational control of mortgage lending standards and enforced standards that permanently removed 1/3 of former borrowers from the market.
At a macro level???
Macro in the sense that someone could infer a housing policy based on macro level events, like PE owned housing tracts.
I think the most powerful source of inference in the policy shift that I have is the charts of credit scores, loan size, and home values on new mortgages at Fannie Mae.
I don't understand. What is your question?
Unfamiliar with the Lucas critique. Pls help.
The Lucas critique is just that if something is a policy goal, then it ceases to be empirical evidence for anything else.
For example, there may be a lot of things that used to correlate with inflaiton. If those things don't have a correlation with inflation any more, it's just because the Fed policy framework removed variance from inflation. That thing may still be inflationary, all else equal, but Fed policy is designed to change other variables in order to counter the inflation.
Kevin, you and are on the same page about a need for a sound national housing policy. Our perspectives come from opposite sides of looking at the same facts, yours as an academician and mine as a hands on operator with 55 years of incredibly diverse experience in all real estate asset classes, types and locations. I’m working very diligently to walk in your shoes and understand your work.
But, I’m struggling a bit with inferring housing policy from tool usage (like RMBS, or Fannie and Freddie loan portfolios and loan policies, etc.). Hoping you can help me here.
Is there a simple cause and effect description for the housing policies that you see? How and what would change that paradigm and result in a greater supply and more “affordable” housing?
Are you asking why I claim that the change in lending standards has made homes less affordable?
Maybe? If you are referring to the current standards, no. That’s pretty clear, and they are an overreaction to something borrowers didn’t create. It’s earlier times, I think. Like the ones used by RMBS pools.
If I understand what you're asking, I would say that there was a bit of a lending bubble from 2003-2007. It was probably associated with temporary inflation in aggregate American housing values of up to about 8%. That's pretty sizeable. I wish that 8% boost had been the most important factor in American home values.
I have problems with other analysis. Home prices nearly doubled from the mid-1990s. The credit portion of the boom was insignificant. That is obvious now, since the arrest of those markets did not stop further price appreciation. Conventional analysis usually controls away 80% of factors (including differences between metro areas), then sort of pretends that the few % of a price boost they attribute to credit markets can be inferred to be much larger. They are wrong. Supply constraints created 80% of the rise in prices, unrelated to credit markets, and there was a credit boom on top of that which would have been sizeable in any other context, but is insignificant compared to our other problems.
Also, the rise in homeownership happened almost entirely before the subprime boom and the peak of the price boom. The private securitization originations that had extremely high default rates happened when prices had leveled off and home sales were in free fall. The timing doesn't allow it to be an important causal factor in more than a small portion of price appreciation. And, finally, the CDO boom was entirely after all housing trends had turned south, and has been systematically mischaracterized as an example of leveraged risk-taking when it was really the result of a market place desperately seeking low risk destinations for savings in a market where home equity was increasingly seen as a risky investment.
I write about it somewhat in "Building from the Ground Up" and I try to measure the effect of credit in this paper: https://www.mercatus.org/research/research-papers/reassessing-role-supply-and-demand-housing-bubble-prices
The line beginning with Notwithstanding should have read “Notwithstanding, on the surface… etc…
Kevin, George, my golden retriever editor was helping me and the last post got sent before it was finished. Hoping you can follow the logic.
Is it possible that the “poor housing policies” you are focused on resulted from an ingrained institutional default need to fill any policy vacuum with something, or anything, when the then politicians weren’t interested in or even capable of providing any well thought out policies and goals?
My recollection is that the nascent GOP tea party politicians were solely focused on deregulation and the destruction of extant policies while simultaneously loosing the dogs of unfettered free market capitalism as the best means to solve all problems.
I think there is something to that.
Additionally they benchmarked to crisis and poverty, so they only supported unfettered free markets where that would lead to liquidation and unsustainably low market prices.
In the end, what that meant is that their only principle was crisis and poverty. And, it was easy enough for them to tell themselves that where markets wouldn't have led to liquidation and panic, it was because of the federal reserve, Fannie & Freddie, etc.
The Federal Reserve was the wild card that could be used to assert that any non-crisis outcome was the result of non-market interventions. This is still common today.
I’m thinking that “benchmarking to crisis and poverty leading to liquidation and unsustainable low prices” aspect is baked into unfettered capitalism by default. I see it as a consequence not a goal. Ascribing such a dark premise as an intentional act in public policy is unimaginable. However, not minimally footnoting it as a possible outcome is imaginable.
Part of my thinking stems from the unintentional but very real black swan risks that live in long tails. Those risk become increasingly likely as systemic underwriting pressures are increased and unintended consequences from loosely related exogenous variables arise. (Say in a heavily reinsured RMBS pool of lower quality loans, coupled with an increasingly skittish investor market that results in the disappearance of some or most of each of the individually rated tranche’s resale liquidity. The inherent liquidity assumption in each and all pools and tranches is that a relatively inexhaustible supply of new or other pools and investors is always available, which of course was and is a fallacy.
Notwithstanding that liquidityn the surface a tranche or pool liquidity failure could be ascribed to your observation. Looking deeper shows that it isn’t; the failure was because of underwriting aggressiveness and market forces. Neither are policy decisions, unless one can conclude that the mere existence of the securitization tools and underwriting rules are the policies. If so, an effective policy determination would be made with each individual pool much like a theoretical adhocracy would work.
It's the Lucas critique.
If there are 10 types of securities, ranging from least to most risky, and public policy becomes benchmarked to some sort of failure, then the most risky securities will fail. But, you can't then conclude that their riskiness was the cause of the failure. The policy was the cause of the failure.
There were certainly ill-advised securities. And, they were what failed. I'm not sure if their failure was inevitable. And, in fact, I would argue that the crisis and the recession were so deep because the amount of economic contraction required to make them fail was much greater than what anyone assumed. The recession was Great because there actually weren't a lot of systemically important securities ready to fail at the drop of a hat.
I wonder if it would seem that way if causing their failure hadn’t become a policy goal.
If Kevin Erdmann is going to write insightfully and reasonably about housing and mortgage issues, how does he ever expect to find an audience?