Monday, February 20, 2017

Housing: Part 208 - Underwriting or money?

Here is a graph of the values of AAA rated securities on CDOs in 2006-2011.

Source
Here also is a graph of NGDP growth in 2006-2011.

So, was it underwriting that led to these changing values?  Or was it monetary policy?  Did the rating agencies that gave these securities AAA ratings err in their estimation of the underwriting standards, or did they err in their forecasts of monetary policy?


Source
Those early 2006 pools were at the height of prices and of mortgage issuance, yet they nearly regained face value even after the recession and the severe collapse in collateral values.

Re-imagine these graphs if the Fed had stabilized GDP growth in the summer of 2008, or the spring, or in late 2007, then re-consider the question of whether the outcomes for these securities were more due to underwriting or to monetary policy. If you are tempted to respond that earlier stabilization would only have encouraged more recklessness, remind yourself that by the time these securities were trading below full value, origination markets for private securitizations were effectively dead.  That ship had sailed.

If we count the closing down of mortgage issuance to middle and lower-middle portions of the market that continued after GSE conservatorship, literally years worth of destabilizing policies were implemented after new issues for the private securitization market died.  The reversal of any of those policies that happened after these securities started trading at discounts would have decreased the damage done.

At the September 2007 FOMC meeting, Bernanke notes in The Courage to Act (p. 162):
As in August, we again discussed the issue of moral hazard – the notion, in this context, that we should refrain from helping the economy with lower interest rates because that would simultaneously let investors who had misjudged risk off the hook. Richard Fisher warned that too large a rate cut would be giving in to a “siren call” to “indulge rather than discipline risky financial behavior.” But, given the rising threat to the broader economy, most members, including myself, Don Kohn, Tim Geithner, Janet Yellen, and Mishkin, had lost patience with this argument. “As the central bank, we have a responsibility to help markets function normally and to promote economic stability broadly speaking,” I said.
In the end, the Richard Fishers' won and the rest of us lost.  The oddity is that it seems to me that the consensus, even in hindsight, remains with Fisher.

17 comments:

  1. Kevin:

    This argument and the market monetarist argument generally assumes that the central bank (the Fed, the BoJ, the ECB, the PBoC) can indefinitely support whatever growth in debt & leverage (especially private sector leverage) is necessary to maintain trend NGDP growth. At some point, as Japan found out and that the US and China are about to find out, this model doesn't work. At least not without massive and today impossible-to-sustain financial repression.


    Howard Marks at Oaktree Capital likes to say that markets abhor certainty. I think he's right. If the markets are certain that the central bank will do whatever is necessary to maintain trend NGDP at some high and stable level, say 5%, then the private sector will add leverage upon leverage to take advantage of that stability until (as in 2007 in the US) the system breaks.

    IMHO, by far the best book on this topic - how we got here, why we can't stabilize here, and where we are going or ought to go - is Adair Turner's Between Debt and the Devil. Independent of one's opinion about helicopter money (Adair's solution), Adair explains why the debt upon debt approach that backstops growth works great until it doesn't and then the system breaks catastrophically. I suspect Michael Pettis would agree with that concept regarding China's monetary management and outlook as well.

    Please read Adair's book and think about the "markets abhor certainty" comment in terms of the standard market monetarist criticism of the Fed during the housing / mortgage meltdown. It's technically true that the Fed never runs out of ammunition. It's also technically true that leverage cannot increase forever. The more certainty there is about the ammunition, the more likely economy-wide leverage will increase to take advantage of it. At some point the system breaks. Saying the Fed could have prevented the break in 2007 is not the same as saying the Fed can forever prevent an increasingly leveraged system from breaking and breaking from an even higher level.

    ReplyDelete
    Replies
    1. Thanks for your comment and for the reading suggestion. However, I have recently been working on this topic in the book, and I think there are many problems with this point of view. I think one of the major sources of policy error has been this idea that debt rises with sentiment. This is wrong. In the recent crisis, rising leverage was the first sign of monetary policy that was too tight and was related to the bust, in my opinion. The rise of CDOs was a sign of the breakdown. These ideas about debt have us so twisted up in confusion that we took a rabid demand for AAA securities as a sign of recklessness.

      On a more secular level, debt is also not the result of loose money or deregulated credit. It is the result of financial repression - the obstacles that prevent capital from being allocated to Closed Access real estate. This increases the value of real estate there, and it does correlate with economic sentiment, because the price of CA real estate reflects a claim on the repressed access to future economic activity, through the ownership of a politically restricted asset. Tellingly, leverage actually tends to be low in CA cities, in terms of debt/asset levels, but it is high in terms of debt/income levels. This is because the value of that real estate is based on the expectation that financial repression (limited housing) will boost the future incomes of future CA residents.

      I would agree that debt is a problem. But, fundamentally, it has little to do with monetary policy. It has to do with the tax code and with housing repression.

      Also, I have developed a rule of thumb that any book about finance that references demons or devils in the title is suspect. But, that's a minor quibble. I will check it out. :-)

      Delete
    2. Adair Turner ties excess private sector leverage to the perceived (or government-supported) stability of existing real estate (as opposed to new construction) in a way that is consistent with your Closed Access cities hypothesis.

      The issue I'm raising is much broader. Greece dug itself into an economic abyss in the first decade of this century over social spending tied to the perceived stability conferred by its membership in the Eurozone. Closed Access cities, residential construction, and financial repression had nothing to do with it. China is digging itself into an economic abyss over fixed investment and excess debt tied to the perceived stability of its central bank and sort-of centrally planned economy. There is plenty of financial repression but there aren't any Closed Access cities. In the case of Spain, it was exactly the opposite of Closed Access cities. It was build, build, build and build some more, all residential and, like Greece, not a hint of financial repression. In Ireland it was residential construction, banking, and no financial repression. In Iceland it was international banking but not residential construction or financial repression. A generation before all of this it was Japan and across-the-board stability (monetary, fiscal, regulatory, exports, currency, etc.) until real estate prices reached the stratosphere and then it was 1989 and debt, debt, and more debt.

      The point is that while localized risks of instability accompany extremes in Closed Access cities in the US, the bigger and much more systemic risk is from excess debt resulting from perceived stability due to monetary policy generally, not the tax code, not housing repression, and not financial repression. This is much more akin to Minsky's Financial Instability Hypothesis than your Closed Access cities hypothesis.

      Whatever stability the monetary, fiscal, mortgage, housing, or regulatory authorities could have put in place in 2007 to avoid the financial crisis once the debt was in place would have only deferred the crisis because the stabilizing influence (e.g., Fed pegging NGDPLT) would have translated directly into more debt and more leverage. The problem is that the incentives to add debt and increase leverage during periods of stability are part of the system. Closed Access cities is a valid explanation in a limited sense but it does not characterize the broader situation - stability is destabilizing (as Minsky said) or that markets abhor certainty (as Marks said). Houston, for example, is about as open as an Open Access city can get. And every time oil prices plunge, so do real estate prices. Nothing to do with monetary policy, the tax code, or housing repression. A lot to do with leveraged asset prices tied to some locally crucial component of stability.

      The systemic issue is the inherent and non-linear instability of a leveraged system and the incentives for members of this system to engage in actions that may be individually beneficial but collectively destructive. As long as the markets are stable, debt and leverage are rewarded. At some unknowable point, the debt and leverage become excessive, the system tips over, and the demands for government-enforced stability grow. This can't go on forever. It's why so many economists want the entire banking system to operate at lower leverage ratios, mortgages to be issued with less debt and more equity, mortgages to have more countercyclical elements (like shared appreciation / depreciation), government bonds to have more countercyclical elements (like GDP-linked interest rates), and so on. It's a much, much bigger issue than Closed Access vs. Open Access cities.

      Delete
    3. Thanks for bringing these issues up. I do need to be able to address them.

      I don't claim that Closed Access is the only problem that can lead to debt or instability.

      And, you didn't mention any of the countries that, as of 2006, were most like the US in terms of economic development and housing market context - UK, Canada, Australia, New Zealand. So far they are going on forever. It simply cannot be argued that we are better off for having cut NGDP growth more than they did in 2008. As far as I'm concerned, when everyone can cite polls from those countries where their residents claim that in hindsight, they wish they had had our crisis, then you can all cite Minsky to me whenever you want. It's 2017. Hunker down. It might be a long wait.

      I don't disagree that debt is destabilizing. I just disagree that stable growth leads to leverage. Think of the full stable of assets at a given time. This ownership is divided between debt ownership, where cash flows are pre-determined, and equity ownership, where income is from residual cash flows. Now, imagine from some starting point that owners become complacent about stability. Explain to me how this leads to a shift in that ownership structure to more debt ownership and less equity ownership. If savers are seeking risk, then why would there be fewer equity owners as a result?

      Delete
    4. The premise of your challenge in the final paragraph assumes that stability means the "the full stable of assets at a given time" is growing in line with stable growth in GDP at some constant debt/equity ownership structure. It is not.

      Whether we're narrowly talking about residential real estate in Closed Access cities or real assets subject to leverage more generally, the issue is that the "full stable of assets" is not restricted to the underlying physical asset or the initial debt/equity structure.

      An entire universe of assets can be created from that underlying physical asset. First lien whole loans and various second lien loans become RMBS or CMBS, which become CDOs, which become CDOs-squared, which spawn synthetic CDOs, which are rehypothecated, and on and on so long as there are willing sources of capital to support the structure.

      Overlay this structure on a system where the apparently stable underlying asset looks like a sure thing to the highly leveraged borrower. In financial terms it is equivalent to an underpriced call option compliments of the residual guarantor, i.e., the taxpayer. The highly leveraged borrower gets a loan from a highly leveraged bank that sells the mortgage to a highly leveraged GSE that sells the MBS to a highly leveraged shadow bank that sells tranches to an entire suite of investors - many of whom embed additional leverage. All of the assets in this structure are as real as the single underlying asset and all or almost all employ leverage.

      In terms of residential real estate, the shift in ownership to more debt and less equity has been going on for a long time. Two generations ago, the norm was 20% down. Allowable PITI/income was 28% and allowable total debt (including student loans, car loans, credit cards, etc.) was 36%. Now, that 20% down is in single digits and 28/36 is well into the 40s.

      Then, one generation ago came all of the derivatives described above plus plenty more. The first level pre-determined cash flows and pre-determined debt/equity ownership is now the basis for a system that incentivizes leveraging that debt (and in many cases, the equity too) as much as short-term stability permits. That's where the shift in ownership structure to more debt and less equity shows up and that's why the more leveraged the system, the more destabilizing stability becomes.

      And to the extent it hasn't led to an implosion in Canada (probably starting with Vancouver), Australia, New Zealand, or the UK, give it time. Consistent with what I said earlier, the central banks of these countries have unlimited ammunition but markets do not permit unlimited leverage. IMO, that's the flaw in NGDPLT or similar stabilizing measures - they incentivize unlimited leverage and ultimately the system breaks.

      Closed Access cities is a very big problem for all the reasons you have described in your blog but it is not the overarching issue. The overarching issue is excess debt, excess leverage, and the demands for central bank policies that permit the debt and leverage to grow forever. Until it can't grow forever. Then the system breaks.

      Delete
    5. MBSs and CDOs cannot create leverage. In fact, they convert some debt to equity. If a bank has 1 billion in mortgages and they securitize them, they are deleveraging. Now, there are, say 950 million in new creditors who have more certain cash flows and lower coupons than the original mortgages had, and someone has a 50 million equity position in the pool. In the aggregate, they reduce leverage.

      Delete
  2. Excellent blogging and commentary.

    I do think Adair Turner's perspectives are very worthwhile to ponder, and not inconsistent necessarily with Kevin Erdmann's.

    Turner asks a reasonable question: Instead of fiscal deficits, why not helicopter money?

    Turner also ponders the relationship between property, money creation and financial institutions. I think Turner would agree that property zoning plays a role in the huge amounts of money being soaked up by extant real estate.

    Switching topics, I encourage both Erdmann and Chernoff to read:

    https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr541.pdf

    This Fed paper, which was later published in the Journal of Money, Credit and Finance (considered a very serious platform) posits large trade deficits cause housing "bubbles," as foreign capital pours into limited real estate.

    Side note to Chernoff: A few years back, the PBoC simply bought a lot of bad loans from their banking system. As the PBoC is below its inflation target anyway, they seem to have this option. The China banking system is probably solid as a rock, as they can offload bad loans on the central bank. This is not my plan, but it is a system that works and actually makes sense as long as the amount of bad loans is kept in check.

    Side note to Chernoff and Erdmann: Turner discusses Japan the BoJ and debt. As both of you probably know, the BoJ has been buying so much national debt, that the Japan government will own the bulk of its debt by 2020.

    I call this Escherian economics. I do not know what to make of charts showing the Japan government to be heavily leveraged. They owe the money to themselves.

    Worth noting, evidently apartment development is easy in Tokyo (still a growing city), and apartment values steady.

    At the risk of being dramatic I think Erdmann has a shot at some sort of unified field theory for modern-day economics, touching upon monetary policy, real estate policy (zoning), money creation and debt and maybe trade deficits results tossed in.

    It is a sad testament on orthodox macroeconomics that the profession insists at barking up the wrong trees, let alone being wrong on some of the fundamentals.

    ReplyDelete
    Replies
    1. "At the risk of being dramatic I think Erdmann has a shot at some sort of unified field theory for modern-day economics"

      I'm afraid the subject matter demands it and that I am not qualified to do it. But, if not me, who?

      Delete
  3. "Richard Fisher warned that too large a rate cut would be giving in to a “siren call” to “indulge rather than discipline risky financial behavior.”

    Gadzooks.

    Yeah, you know, like mostly middle-class people who buy houses when they see rents relentlessly rising, or institutional investors who bought AAA-rated mortgage-bcked bonds.

    You let people get away with risky behavior like that, and the next thing you know, people start buying stocks on Wall Street--or even risking capital to start up a business or build housing.

    Who knows where this could all lead?

    An extended and deep recession is what people need to keep them in line.



    ReplyDelete
    Replies
    1. It's a deep seated notion we have, that our lack of humility must be our undoing. It certainly pre-dates capitalism, though it apparently hasn't been suppressed by it.

      Delete
    2. A couple of follow-ups.

      Japan is getting to the position where the BoJ will own all the public debt. So long as it can pull this off without a currency crisis or massive capital outflows, it's sustainable. That's what seems to be happening so it can continue.

      China is years behind Japan in creating a credit boom / deflationary bust but it's on the same track. It's already facing massive capital outflows and pressure on the currency. It can't continue. It also appears that China has wasted a lot more of its domestic fixed investment on assets that cannot service the associated debt than Japan has.

      The connection between current account deficits and real estate prices in places with desirable real estate is solid and a very big deal. Kevin's Closed Access cities argument connects directly to this. Another issue that connects directly to this but gets little or no attention is that the underpinnings of all the free trade models in economics dating back to Smith and Ricardo are based on trade in goods and services, not free capital flows into foreign real estate. Or, maybe in a more technically accurate phrasing, the US has a comparative advantage in property rights, the rule of law, and really nice coastal cities. Of course, a lot of Americans inside and outside those really nice coastal cities don't think they or anyone else in the US benefits from this kind of trade.

      Finally, MBS and CDOs absolutely increase leverage in the overall financial system. The example Kevin provides increases leverage in the system even if the leverage for the original borrower buying the real estate doesn't change.

      MBS and CDOs also allow more capital (debt and equity) to be directed towards real estate than would otherwise be the case, putting upward pressure on real estate prices. Here's a thought experiment: Estimate the long-run equilibrium for housing (house size, % of GDP, % of household assets, ownership vs. rental, etc.) in the absence of the GSEs, MBS and CDOs, shadow banking and SIVs, etc. I would argue that we'd have very roughly the same number of houses and very roughly the same ownership vs. rental percentages but that the houses would be much, much smaller and much, much less expensive. They would represent a much smaller percent of household wealth and a much smaller percent of financial system assets. Take away tax preferences (mortgage interest deduction, capital gains exclusion, etc.) and the prices go down some more but everything else stays roughly the same. That system, with one level of leverage (the initial loan) at much lower LTV and DTI and no MBS or CDOs would be a lot more stable. We'd still have Closed Access cities, however.

      Delete
    3. I find that the scale of the subsidy to homeownership from tax benefits vastly outweighs anything due to MBSs. Many, many times larger. And in this context, credit creates access. High end real estate sells for much higher multiples because of more universal access to ownership and mostly because of more universal access to tax benefits. The problem in the Closed Access cities during the boom was that marginal expansion of credit to high income households that didn't meet conventional standards meant that lower priced Closed Access real estate could be bid up to higher prices because the new buyers could take advantage of tax benefits. There was a sort of perfect storm - tax benefits, deprived supply, and a shift in credit standards. Out of these three things, it's the taxes and the supply problem that need to be fixed. Access is a public good.

      In this case, access appears to have been different than what most people believe. There wasn't an increase in ownership among households with marginal incomes. Instead, access raised prices in low end CA neighborhoods to a more efficient price, inducing an increase in the migration pattern that CA causes, where households segregate by economic potential. The effect of credit expansion was to create a more efficient housing market, but with Closed Access, efficiency kind of means more people are excluded more efficiently. Cutting off access, however, is the opposite of the solution, and making debt the centerpiece of our worries is a step in that direction. In fact, that is what we have been doing for a decade, to tragic effect.

      Delete
  4. Henry C--

    Capital outflows slowed down from China in Jan. They still have trillions of dollars of reserves.

    Curiously China has the same problem as US...in job generating cities a lack of housing due to zoning...

    Their financial system is indestructible, if less efficient than US's...

    Their political leadership is nutty...

    ReplyDelete
  5. I'm afraid the subject matter demands it and that I am not qualified to do it. But, if not me, who?--Kevin Erdmann

    I have been puzzling this. I keep thinking, "Sure I am a smart guy. Probably Kevin is even smarter. But surely, there are hundreds, probably thousands of smart guys thinking about monetary economics, property, etc. I am crazy to have my views?"

    I think the answer is not that I am crazy, at least in this instance.

    I think there is a tremendous amount of power in convention, orthodoxy, professionalism, and careerism.

    Simple example: Surely, property zoning is roughly equal to, say, the minimum wage as a structural impediment.

    Google "property zoning and economics" and or "minimum wage and economics."

    It is not conventional to talk about property zoning. Let alone push-cart vending.

    I suppose there is some class bias in this; Mercatus and Hoover etc. have financial backers, and they have interests in the property-financal system nexus. But even the Berkeley or UT crowd does not talk about property zoning. It is not an orthodox topic at this time.

    And one one anywhere wants to ever breath a word of even perceived opposition to free trade, so the connections between trade deficits and house prices cannot be be aired. You might as well be a Trumpster.

    That is my guess anyway. Either that, or Kevin Erdmann is a nut, and I am his follower.

    ReplyDelete
  6. Excellent post and excellent commentary (all around). My thinking has often led me to wonder about whether a successful period of NGDPLT would lead to its own catastrophe due to over-leverage. Eventually, it probably would. Proper regulations could potentially manage to stave off the crisis for a long time though. LTV requirements, ending the tax deduction for interest, increasing equity requirements for banks and other things like that. But of course entrepreneurs will always seek out ways to circumvent the regulations and regulators wil eventually fall behind. I think the 2007-2009 crisis was avoidable (at least delayable). And I think the eventual recession could have been smaller. No system or solution will ever be the final, perfect one though. But every loan has a lender and debtor, and the lenders have an interest in making mostly good loans.

    ReplyDelete
    Replies
    1. I think I will do a post on this. Others have had similar responses. I don't think it is clear that leverage would increase at all.

      Delete
  7. Note ABX 06-1 typically means the loans were 2nd-half 2005 vintage. Actual 2006 vintage looks far worse on your chart. That said, even to say the AAA 06-1 tranche 'nearly regained face value' could be seen as a questionable way to summarize a supposedly AAA asset that settled down at 90 cents i.e. lost 10 cents on the dollar. (It's possible a loan-level analysis would help your case in a way this one does not? Or, it might not)

    Relatedly, later on you say MBS/CDO do not increase leverage. No not in the initial act of the structuring itself, but that ignores the fact that at least some target end buyers could then buy the resulting, say, AAA piece with a lot of implicit leverage due to its being AAA. The ABS CDO market then repeated the process. It also ignores the effect of the CDS market (including ABX) which allowed for leverage that was, in principle, only limited by whatever swap/counterparty agreement the buyer had with the seller, again often ratings-based.

    I'm generally sympathetic to arguments for more accommodative monetary policy but one can't ignore that leverage, exploiting a clunky regulatory arbitrage around 'ratings' (and reverse-engineering bad rating-agency models), fed these assets. Or that much shoddy underwriting if not outright fraud occurred as (probably at least in part) a result. Those were fundamental problems with this market independent of monetary policy that, even if they could've been papered over for a while?, were bound to burst sooner or later. (Some would actually say they *weren't* independent of mon. policy but in the opposite direction - on grounds that demand for these assets was fed by early-00s rates that were 'too low', though I'm not sure I'd go that far...)

    ReplyDelete