Home > Federal Reserve, Trading > Are CMBS Strains Coming To The Fore?

Are CMBS Strains Coming To The Fore?

One thing is certain these days. By sitting on my hands, I’m missing a lot of tradable swings! I am not too disturbed by this fact, since I am probably almost as likely to trade them poorly as well, but we are most definitely in a trader’s market. Whatever your sense of the current economic environment makes you bullish or bearish, the value of corporate claims is not changing weekly by 140 S&P points. From the October 4th lows, stocks closed higher today by 12.4% (from the October 3rd closing low, the gain is only 9.8%!).

Are people really investing in this kind of crazy market? It scares me to death, and I’m a professional! (Then again, with all of the effort that officialdom has put into soothing us, perhaps it’s only the professionals who are really scared.)

Intraday, the Dow erased its loss for the year before falling back. That sounds impressive until you remember that it was up 10% on the year as recently as July 21st. Volumes were fairly low and the Vix declined but remained above 31. 10-year Treasury yields rose to 2.21%, with 10-year TIPS yields at a hearty 0.23%. Copper and precious metals rallied; grains and energy declined.

The FOMC minutes from the September meeting were released today. There were some interesting items in there, aside from the somewhat crazy assertion that “most FOMC officials said inflation appeared to have moderated.” (see Chart)

Moderating inflation, according to the Fed. This is core CPI, year-on-year.

Bad forecasts and weak monetary policy judgments from the Fed no longer surprise me, sadly, but two notes made me raise my eyebrows a little. Here is the first:

However, some others noted that a recent change in deposit insurance assessments had the effect of significantly reducing the net return that many banks receive from holding reserve balances.

This statement occurred in a discussion about whether interest on excess reserves (IOER) ought to be cut as a stimulative policy measure in order to spur bank lending (apparently they don’t believe the hype about how there’s just no loan demand). It is interesting because it implies that IOER has already been cut, effectively, and this in turn makes me more nervous about the recent rise in M2. For some time I’ve been suggesting that if the Fed wants to really add liquidity, it needs to eliminate IOER and that would tend to flush reserves back into the system. This snippet suggests that something similar has already happened. And that in turn increases the chances that the recent M2 rise is more than just a European deposit-flight one-off. I wasn’t aware of the change in deposit insurance assessments and I don’t know how large the change is, but the FOMC members appeared to think that it was enough to cause a meaningful change in the effective IOER realized by member banks.

The second item would have slipped past me if I hadn’t just seen an interesting chart on the Fed’s own home page. Here is the item from the minutes:

In secured funding markets, term financing reportedly remained readily available for both domestic and European financial institutions through repurchase agreements backed by Treasury and agency collateral. However, some strains emerged late in the intermeeting period in the market for repurchase agreements backed by lower-quality, nontraditional collateral.

And here is the chart:

Borrowing strains are appearing in ABS. Source: Federal Reserve

Lower-quality, nontraditional collateral like AA asset-backed lending, perhaps? I am not tied into the ABS market but I wonder if this isn’t a warning sign for something bigger, like strains in the CMBS market for example (recall that the heads of the CMBS groups at Goldman and Citigroup both resigned the week of … July 28th, and I noted it here. July 28th marked the point just before the first spike on this chart).

Some observers have been looking for an echo-blowup from the CMBS market for a long time after the residential MBS market took heavy fire in 2007-09. With everything that has been going on in Europe I haven’t heard any developments recently, but this FOMC minute and this chart suggest that we should pay more attention. Owners of CMBS-backed paper ought to be looking quite carefully at their exposures not just to the underlying assets but also to the availability of financing on structure.

Tomorrow, a quiet data week continues with only Initial Claims (Consensus: 405k from 401k) on tap. Equities need to break higher fairly quickly, or the bearish crowd will team up with the range-trading crowd and push prices lower again. I’m rather in that camp myself. I am not concomitantly bullish on bonds, though. I would love to have the opportunity to get short 1.80% on 10-year notes again, because I missed it the first time (I judged it too early to be fading the strong seasonal pattern and have missed a 35bp one-way selloff since then), but I don’t think that’s going to happen. I’d be happy to sell some around 2.1% and scale into a short on a further rally, but I wonder if I will even get a chance to do that. I will probably use RYJAX, which is a single-inverse fund that doesn’t have the return drag that a leveraged-inverse fund gets from volatility.

And come to think of it, maybe I’ll start doing that right away in case I’m wrong about stocks!

Categories: Federal Reserve, Trading
  1. October 14, 2011 at 7:21 pm

    re: the recent rise in M2 and possible inflation concerns.

    In response to another blog (about China and M2) I had a look at US M2 growth and the CPI to see if there was any link. There is nothing to see there. I wrote it up anyway:


  2. October 14, 2011 at 11:12 pm

    I appreciate the link. It looks like you did the analysis with headline inflation, though. The noise in headline inflation due to stochastic fluctuations in energy prices will completely swamp the signal (which is one reason the Fed doesn’t watch headline inflation)…no surprise to me you didn’t find anything. There is a good fit (as econometrica goes) with core inflation, with a reasonable lag, if you control for private debt (which dampens inflation).

    Also keep in mind that in the 1990s and 2000s, when inflation was low and fairly stable, you had little signal and so your signal:noise ratio makes it hard to find much. You can still find the relationship there if you control for the proper variables.

    • October 15, 2011 at 12:52 am

      ok. I ran it all again using CPI (all items less food and energy) from the BLS tables: http://data.bls.gov/pdq/querytool.jsp?survey=cu

      Still nothing to see: slight positive correlation (0.1) but still poor r squared (0.04). I’m not sure what sort of econometric method you need to apply to tease a desired result out of this but on the face of it there is just nothing to see. I’ll put the revised data up as well.

      I’ve not controlled for private debt — but then again I never hear or read qualifiers about the need to do this when people discuss M2 and inflation.

      • October 15, 2011 at 8:04 am

        I know you’ve never read about that…I am pretty sure I’m the first person to introduce that factor, which you need if you are to predict both the 1980s and the 1990s-2000s. You can fit one or the other but not both. Our model incorporates the FX value of the dollar, money supply (thus controlling for ex-US money supply, which also affects US inflation of course), a trend variable for inflation (which you obviously need!), and the ratio of private debt to total debt. I think this last factor needs further refinement; it’s probably not linear – but it works well enough. Anyway, these factors don’t have big collinearity problems, like Fama/French pointed out with things like money and industrial production, and they explain the level of core inflation quite well. (There are lags in each variable, which we take as static lags but really should be distributed lags).

        We also have a model that forecasts core-ex-housing and then adds a separate forecast for housing, derived from lagged inventory levels. That model uses the same overall inputs but slightly different parameters.

        And now I’ve told you WAY more about our proprietary models than I ought to have! 🙂

        Keep on regressin’ pal!

  3. October 15, 2011 at 4:12 pm

    But now we’re talking multivariate analysis. The thing that got me started on this (as above it was initially a blog discussing China inflation and M2 but since I read your blog and know you have discussed M2 thought I would come here to a chat 🙂 ) is the suggestion that M2 growth implies something w.r.t inflation, either inflation now or future inflation. So I am explicitly testing verbal claims that get made. If you have to torture the data to get it to give up its secrets then there is no explicit direct relationship. Now it is fair enough you are saying that you guys model core inflation quite differently (in this respect we are talking apples and oranges), but nevertheless a rise in M2 growth in itself should not make you nervous.

    Incidentally, w.r.t. private debt I would imagine it is the velocity of private debt (1st derivative) that may have an impact on inflation via its effect on aggregate demand.

  4. October 15, 2011 at 9:28 pm

    Well, it’s multivariate analysis but that doesn’t mean there aren’t coefficients on the variables! And the coefficient on M2 is significant and positive. It’s when you have undiscovered variables (an underspecified model) or collinear ones that you need to be skeptical about the coefficients.

    I’m not saying our model is perfect; indeed, since we are outside of ranges that the model was built on, you’re extrapolating effects and that’s never a good idea. But I think it’s a pretty decent model, and like most models of monetary effects it shows a relationship between the quantity of money and inflation.

    Of course, there is always room for reasonable discussion! And thanks for contributing.

  5. Jim H.
    October 17, 2011 at 10:21 am

    The noise in headline inflation due to stochastic fluctuations in energy prices will completely swamp the signal (which is one reason the Fed doesn’t watch headline inflation)…

    Even if that’s true, I still can’t embrace the ‘that don’t count’ core CPI.

    Couldn’t we all just use the Cleveland Fed median CPI and get along?

  1. No trackbacks yet.

Leave a Reply

%d bloggers like this: