« Creation Science, oil drilling, naked shorts and constitutionally protected stock fraud | John Hempton's Blog | Australia – the lucky but unbalanced country »
The second derivative is bad
I have been firmly in the “second derivative is good” camp for some time. Green shoots were few and far between – but the economy no longer appeared to be in free-fall. When the free-fall stopped it was time to buy equities – and whilst it was not time to ease up on the looser monetary and fiscal policies – it may have been sensible to limit them somewhere near the levels that they now are.
The data I considered most persuasive was the delinquency data at Fannie and Freddie. It gets worse every month, but until the last data point it was getting worse at a decreasing rate (especially if you adjusted for the foreclosure moratoriums they implemented).
Today I am more worried. My favourite data point (rate of increase of Freddie Mac delinquency) has deteriorated – especially in their insured portfolio. Its not sharp deterioration – and it is possible – even likely – that Freddie Mac will have end credit losses considerably lower than the bears anticipate. But as a second derivative bull I am feeling just that little bit less certain.
Contra: the usually bearish calculated risk has a fairly good data point here.
John
For the real masochists – here is the monthly data from Freddie Mac. The brilliant interest rate management I identified recently has continued albeit not with the panache of the previous month.
Finally BondInvestor has not contacted me as requested. I really would appreciate it.
Read more at John Hempton's Weblog
Advertisement
If Mark Thoma is reading Hempton, there must be some there there, but ---
If Hempton expects us to follow his argument (rather than treat it as a simple, unsupported assertion), it's incumbent upon him to provide a few "numbers" and maybe even a graph rather than just a link to FME's Monthly Volume Statement -- and then, follow up with a bit of analysis of said "figures."
So mortgage delinquencies are rising. Why would we expect anything different at this stage of the recession and with unemployment expected to continue to increase? Reporting the obvious and expected doesn't qualify as analysis.
Why do rising delinquency rates make the "green shoots" meme questionable?
June 27, 2009 1:42 AM | Reply | Permalink
Well, second derivatives are notoriously noisy and I wouldn't put too much stock in one data point. (You are talking about one data point, right?).
Anyway I went to the Freddie Mac data page and I couldn't find any columns with a header that had anything to do with "delinquency". I think that's why people hate finance-- nobody knows what in the hell you guys are talking about. Besides, it's boring.
June 27, 2009 9:26 AM | Reply | Permalink
"Second derivatives."
Uh, would that be like, uh -- second derivatives in calculus -- or like, uh, second derivatives in financial geek-speak*?
* Where the "noise" comes not from the data but from the wannabe speaker's elevated sense of his own importance.
June 27, 2009 12:34 PM | Reply | Permalink
I didn't know there was a difference, but in the post it sounded like he was talking about calculus:
"...but until the last data point it was getting worse at a decreasing rate..."
So here I think the noise is coming from the data.
June 27, 2009 6:11 PM | Reply | Permalink
That was fascinating reading. Is there a translation to English available?
June 27, 2009 12:15 PM | Reply | Permalink
If I understand Hempton, then I think he is correct. He is saying that maybe the economy is not really starting a rebound. That seems entirely likely to me. Look at what is going on at the grass roots - the majority of businesses are losing money. Automobiles, residential construction, commercial construction, tourism, banking, and all the businesses that pivot around these industries are all losing money with no change in the near future. This can't go on indefinately. I think we are in a false bubble rebound that will deflate.
June 27, 2009 8:23 PM | Reply | Permalink
And the Fed will call it "a little froth."
June 28, 2009 12:04 AM | Reply | Permalink
The lack of comprehension of this post was a function of my writing primarily an investment blog - but having policy/politics readers.
The economy is not improving in the things that matter to people (employment, chance of having your house foreclosed on, chance of having someone come through your shop door or ringing and offering you contract work).
However until recently you could say - with a high degree of certainty - that it was getting worse less fast.
November last year was just free fall. It got worse at an increasing and unpredictable rate. If it gets worse like that there is no plausible way you can estimate how bad it is going to get. This was dire.
But by March for most indicators it was getting worse - but at a rate that was slower than a few months prior and at a rate that was itself slowing down. The "second derivative was good". And yes - calculus was the analogy.
That was great. If the fastball is slowing down from 97 to 95 it will still hurt when it hits you - but at least it is slowing down and if you stand far enough back you are safe.
Whereas if the fast-ball has a rocket attached to it and it is speeding up there is no way of estimating how far back you need to stand to be safe.
The turn in the stock market was not green shoots - they were few and far between. It was simply that the crisis appeared manageable.
I have seen a few indicators lately that have bad second derivatives - ie getting worse at an increasing and unpredictable rate.
I am most interested in SUPER PRIME MORTGAGES and indicators of financially conservative middle America. If the delinquency on 800K loans taken by people with 100K household income in the Inland Empire get worse - then who cares? I have already written them off - as has anyone who thinks clearly.
But if 220K loans by people with 80K household income start going bad that is bad bad news.
The second derivative has been surprisingly good at Freddie Mac (and not quite so good at Fannie).
Its now turned sour.
That is BAD BAD BAD.
Is that enough English for you?
John
June 28, 2009 4:39 AM | Reply | Permalink
Plenty of English but not enough Math.
That bump in the Truck Tonnage curve in May is no bigger than the noise bumps right next to it, so just looking at the graph I'd say it's meaningless.
It's tricky to fit a curve with so few data points but if nothing else one could try a low pass filter. Drawing the graph through all the points the way they did at the link is just misleading.
Also, are you taking the second derivative of one or more of the data columns at the Freddie Mac link? Which columns? You mentioned "delinquency data" I couldn't tell which columns that would rever to.
June 28, 2009 8:12 AM | Reply | Permalink