Brad DeLong on the Employment-to-Population Rate

Brad DeLong offers his musing here on the U.S. employment-to-population rate. As a rough control for demographic factors, he focuses on prime-age workers--those aged between 25-54. And he decomposes prime-age workers by sex. I like this exercise. In fact I've looked at the same data in an earlier post here, making comparisons with Canada. 

DeLong normalizes the E-P rate for males and females to zero in the year 2000. For both sexes, the E-P rate is roughly 5% lower than in 2000. Here is his graph:

In reference to this data, DeLong asks a few questions. Heck, it's the end of the year and I'm in the mood to answer some of them.

(1) If the US economy were operating at its productive potential, the share of 25 to 54-year-olds who are employed ought to be what it was at the start of 2000. Back then there were few visible pressures leading to rising inflation in the economy.
Does anybody disagree with that?

I'm not sure, so I think I'll disagree. It's very hard, I think, to know precisely what constitutes "productive potential." And the use of the word "potential" leads us (possibly incorrectly) to interpret the deviations in the graph above as "cyclical" (mean-reverting) fluctuations. I think that some of the decline in the male E-P rate can be reasonably thought of as being "below potential." I base my assessment on this graph (which I plotted a year ago and uses the 16+ population as a base):

This longer time-series shows a modest secular decline in the E-P ratio in both the U.S. and Canada since 1976. Personally, I think it's unlikely that the local peak in 2000 represents some magical measure of "potential." My hunch is that there are "structural" factors at play here including, but not limited to, things like rising disability rates. Having said this, I also don't believe that the male E-P rate has fully recovered. As I've mentioned before, the current dynamic resembles very much what Canada went through in the 1990s, i.e.,

The idea that the female E-P ratio is far below "potential" is even more misleading, I think. There's something strange going on with U.S. female employment relative to other advanced countries (all which resemble Canada in the diagram below).

Again, it's likely that a part of the post-2008 decline is cyclical. But its even more likely that most of the decline is structural (e.g., changing maternity leave benefits, etc.). My own feeling is that structural issues are better tackled through fiscal (or labor market) policies--not monetary policy.

(3) Even if you think–in spite of the absence of accelerating inflation–that employment in 2000 was above the economy’s long-term sustainable potential, there is no reason to believe that a U.S. economy firing on all cylinders would not have 25-54 employment to population rates–both male and female–back at their 2006 levels, a full 3%-age points–and 4%, 1/25–higher than today.
Does anybody disagree with that?

I'm not going to disagree with this.

(4) The U.S. economy’s convergence towards its potential is very slow: The 25-54 employment-to-population ratio has only risen by 1%-point over the past two years.
Does anybody disagree with that?

I'm not going to disagree with this either. However, my interpretation of this phenomenon is a "structural" one, which I explain here (see also here).

(5) Yet in spite of all these, the Federal Reserve believes that the U.S. economy is now close enough to its productive potential that unless some more things go wrong it is no longer appropriate for it to be buying assets and it will be appropriate for it in a year to start raising interest rates even though inflation is still below its 2%/year target.

Well, I'm not speaking for the Fed here (and remember, there are divergent views within the FOMC), but the consensus view seems to be that inflation is just "temporarily" below target. And the recent FOMC statement makes clear that any rate hike will be made contingent on the incoming data. Moreover, even if a rate hike is in the making, it is likely (in my view) to be described as progressing at a "measured pace," similar to the way it was in 2004.

The only way to square (1) through (4) with (5) is if the Greater Crash of 2008-2009 and the still-ongoing Lesser Depression really have pushed between 2 and 4%-age points of our 25 to 54-year-olds out of the labor force permanently, so that we can never get them back, or at least never get them back without an economy at such high pressure to produce inflation that the Federal Reserve regards as unacceptable.

Yes, I suppose this is just another way of saying that a major component of the decline in the E-P rates reported above are attributable to "structural" factors that are better dealt with (if at all) with fiscal policy.

This may be true.
But it does raise two questions: 

[1] What has made the Federal Reserve so confident that it is true that it is willing to make policy based on it–especially as current inflation is still below the 2%/year target?

Again, I am not speaking for the Fed here. But one argument I often hear is that additions to the Fed's balance sheet have not been very effective, especially in terms of improving the labor market. At the same time, people have expressed some degree of nervousness over "not knowing what they don't know" about operating with such a large balance sheet. The Fed is charting new territory here. The benefits seem small at best, and the risks are not fully known. There is some concern that a low-rate policy may induce a "reaching for yield phenomena," leading to financial instability.

Yes, inflation is still below the 2% target, but not that much below. Does a 1.5% inflation rate warrant a policy rate of 25 basis points? Historical Taylor rules evidently suggest that a higher (but still low) policy rate is desirable in the present circumstances. (Note, once again, this is not necessarily my own view.)

[2] If it is true that the missing 2 to 4%-age points of 25 to 54-year-olds now out of the labor force could not be pulled back in without allowing inflation to rise above it’s 2%/year target, isn’t that an argument for raising the 2%/year target rather than accepting the current 77% 25-54 employment to population ratio as the economy’s limit of potential?

No, I don't think raising the long-run inflation target (to say 3% or 4%) will have any measurable long-run impact on the labor market. A significantly higher inflation tax may even discourage employment (the long-run Philips curve may be positively sloped). If there are things that need "fixing" in the labor market six years into the recovery, they are probably better dealt with directly through labor market policies (like improved maternity leave benefits) or fiscal policies (tax cuts, wage/training subsidies, etc.)

***
Some background reading: Many Moving Parts: A Look Inside the U.S. Labor Market.

Read More

Considerable Time and Patience a Decade Ago

According to USA Today:

Wall Street cheered as the Federal Reserve used a new word — "patient" — to basically let the market know that it isn't in a rush to hike short-term rates next year.

So, "patient" is the new buzzword. In other words, the Fed evidently ran out of patience with "considerable time." 

But just how new are these buzzwords? They're not new at all. Consider this from the December 09, 2003 FOMC statement, for example.

However, with inflation quite low and resource use slack, the Committee believes that policy accommodation can be maintained for a considerable period.

This "considerable period" language was also used in the August 12, September 16, and October 28 FOMC statements leading up to the December statement. The FF target rate at that time was 1%. Headline PCE inflation was running at about 2% (year-over-year), and the unemployment rate was about 5.5%.

Then, at the next FOMC meeting, the Fed switched from "considerable period" to "patient." From the January 28, 2004 FOMC statement: 

With inflation quite low and resource use slack, the Committee believes that it can be patient in removing its policy accommodation.

Note that "inflation quite low" in January 2004 was 2%. The FOMC continued to express "patience" in its March 16 statement. In its May 4 statement, "patience" was replaced with:

At this juncture, with inflation low and resource use slack, the Committee believes that policy accommodation can be removed at a pace that is likely to be measured.

Note that the "with inflation still low" statement now corresponds to PCE inflation running around 2.5%. 

It wasn't until the June 30 statement that the FOMC finally raised the FF target rate by 1/4%. And for the next 17 meetings, the FOMC raised its policy rate by 25 basis points. 

Should the Fed at that time exhibited less patience, both in the the timing and pace of "lift off?" Certainly John Taylor seems to think so. 

And what about the situation today? While the unemployment rate today (5.8%) is not far from where it was eleven years ago, the policy rate is at 1/4% (instead of 1%) and the PCE inflation rate is at 1.5% (instead of 2%). At the risk of oversimplifying, there are basically two views on the matter.

The dovish argument is that with inflation and inflation expectations low (relative to target) and unemployment still elevated somewhat, keeping the policy rate at its floor seems like the right thing to do right now. What this has to do with "considerable time" or "patience," I'm not sure. It is a state-contingent policy. (Adding "considerable time" or "patience" to the statement simply reveals the FOMC's own assessment of the probabilities associated with future states of the world.)

The hawkish argument is that the real economy is basically back to normal, that while inflation and inflation expectations are currently low, this is largely transitory. And in any case, the welfare cost/benefit of 1.5% inflation vs. 2% inflation is virtually nil. So, with inflation and unemployment at close to normal levels, why shouldn't the policy rate also start moving closer to normal levels? (There are also other concerns relating to low interest rates and financial instability--look at what happened the last time we had a "patient" Fed.) 

Stay tuned, folks.

Read More

How to dump and restore a postgresql database

To create a postgresql dump in linux, execute the following commands:

1. Login as root user
2. >su - postgres
3. >pg_dump -U username -W -Z 9 -f /tmp/dump_username.dump.zip -h 127.0.0.1 dbname
4. >password

To restore, execute:

1. >psql -U username -f dump_username.dump dbname

Other ways:

>pg_dump -U <username> -W -Z 9 -f /outfile.dump.zip -h 127.0.0.1 <dbname>
><password>

//to restore
>psql -U <username> -f <infile> <dbname>

Read More

Blurred lenovo laptop display on external LG monitor

Recently I've bought a new Lenovo G50-70 laptop to replace my existing one which only support 8GB nax memory. The new one I've upgraded to 16GB, cool it is running my previous one. But to my disappointment after installing Windows 8.1, I've encountered a seemingly common blurred display issue on external monitor - in my case LG LED. I've also tried it on my new HP LED and worked on the configuration as suggested in various articles available in google. But all failed :-)

It seems like there is no solution until windows release an updated driver for high dpi led, but I found one. Fortunately I have an old AOC LCD monitor and so from the hdmi port in my laptop, I plugged in a converter to serial port that goes to my LCD and the display is way way better :-)

This is how eclipse (an ide, a tool for programmer) it looks like on the old LCD monitor.

See how crisp it is. I'll not post an image of how it looks like in my LG LED because that's why you're here after all, having the same display on your machine.

So I suggest you try plugging an old LCD to your laptop to temporarily solve the display crispiness problem.

Read More

Idealavkastning

Oljefondet klarer ikke å oppnå avkastningen som politikerne har forutsatt. Gapet skal tettes med mer risiko. 

De aller fleste økonomer er samstemt om at avkastningen på kapital er lavere nå enn før. Har du penger på høyrentekonto, kan formuen din i realiteten vil være mindre om ett år fordi rentene ikke lenger klarer å holde tritt med inflasjonen.

Fundamentet til handlingsregelen er en realavkastning på fire prosent. Det er urealistisk i dag, Obligasjonsinvesteringene, som utgjør 40 % av fondet, vil kanskje ikke gi realavkastning av betydning i det hele tatt. Mesteparten av obligasjonsinvesteringene er sikker statsgjeld som gir svært beskjeden renteinntekter for fondet. For å oppnå reel avkastning utover prisstigning må aksjeinvesteringene gjøre jobben. Det er imidlertid grunn til å tro at også disse investeringene vil gi lavere avkastning i årene fremover.

Oljefondets svar på denne utfordringen er å ta mer risiko. Fondet ønsker å øke risiko og avkastning ved å gå inn i unoterte aksjer. Unoterte aksjer kan være vanskelig å kvitte seg med når det røyner på. Investorer flest liker ikke det, derfor gir slike investeringer større avkastning.

Men er det virkelig hensiktsmessig å la avkastningen styre hvor mye risiko fondet bør ta? Det har vært bred enighet om at fondet skal ta moderat risiko. Betyr lav avkastning at vi nå skal gå i en ny retning?

Fondet har et «evighetsperspektiv». Det betyr at fondet i prinsippet tåler mye mer risiko enn vanlige investorer som kanskje ikke ser lenger enn til sin egen pensjonstilværelse. Fondet kan derfor i utgangspunktet ta langt mer risiko enn det gjør i dag. Kanskje burde ikke Oljefondet investere i sikre obligasjoner i det hele tatt?

Problemet er imidlertid at befolkningen følger med på fondets svingninger. Halveres fondets verdi i løpet av noen måneder, kan en regjering kanskje se seg nødt til å selge risikable investeringer til spottpris for å berge seg politisk.

En strategi for å høste likviditetsgevinster fra markedets kortsiktige spekulanter kan derfor kollapse lenge før inntektene kommer. Av denne grunn er de fleste enige om at fondet reelt sett ikke har frihet til å operere som «evighetsinvestor». Risiko er viktig for fondet. Derfor investerer fondet en betydelig del av verdiene i obligasjoner.

Det kan godt hende at fondet i dag tar for mye eller forlite risiko, men endringer i risikoprofilen bør være uavhengig av forventet avkastning. Det er grunn til å tro at fire prosent avkastning er uoppnåelig. Å trykke enda mer på gassen for å nå et avkastningsmål fremstår imidlertid ikke som en hensiktsmessig løsning på problemet.

Årsaken til lavere avkastning kan ligge i en velkjent økonomisk mekanisme. Uten innovasjon er det slik at mer kapital per arbeider gir lavere avkastningen på kapitalen. Oljefondets fallende avkastning kan altså være et resultat av at innovasjonstakten ikke lenger klarer å holde følge med kapitalveksten. Det kan rett og slett være at stadig mer kapital jakter på et begrenset utvalg med gode prosjekter.

Det er i så fall lite Finansdepartementet eller Norges Bank kan gjøre med saken. Det kan godt hende det finnes saklige argumenter for at Oljefondet bør ta mer risiko, men lav forventet avkastning er ikke ett av dem.

Read More

afl-fuzz: nobody expects CDATA sections in XML

I made a very explicit, pragmatic design decision with afl-fuzz: for performance and reliability reasons, I did not want to get into static analysis or symbolic execution to understand what the program is actually doing with the data we are feeding to it. The basic algorithm for the fuzzer can be just summed up as randomly mutating the input files, and gently nudging the process toward new state transitions discovered in the targeted binary. That discovery part is done with the help of lightweight and extremely simple instrumentation injected by the compiler.

I had a working theory that this would make the fuzzer a bit smarter than a potato, but I wasn't expecting any fireworks. So, when the algorithm managed to not only find some useful real-world bugs, but to successfully synthesize a JPEG file out of nothing, I was genuinely surprised by the outcome.

Of course, while it was an interesting result, it wasn't an impossible one. In the end, the fuzzer simply managed to wiggle its way through a long and winding sequence of conditionals that operated on individual bytes, making them well-suited for the guided brute-force approach. What seemed perfectly clear, though, is that the algorithm wouldn't be able to get past "atomic", large-search-space checks such as:

if (strcmp(header.magic_password, "h4ck3d by p1gZ")) goto terminate_now;

...or:

if (header.magic_value == 0x12345678) goto terminate_now;

This constraint made the tool less useful for properly exploring extremely verbose, human-readable formats such as HTML or JavaScript.

Some doubts started to set in when afl-fuzz effortlessly pulled out four-byte magic values and synthesized ELF files when testing programs such as objdump or file. As I later found out, this particular example is often used as a benchmark for complex static analysis or symbolic execution frameworks.

But still, guessing four bytes could have been just a happy accident. With fast targets, the fuzzer can pull off billions of execs per day on a single machine, so it could have been dumb luck.

(As an aside: to deal with strings, I had this very speculative idea of special-casing memory comparison functions such as strcmp() and memcmp() by replacing them with non-optimized versions that can be instrumented easily. I have one simple demo of that principle bundled with the fuzzer in experimental/instrumented_cmp/, but I never got around to actually implementing it in the fuzzer itself.)

Anyway, nothing quite prepared me for what the recent versions were capable of doing with libxml2. I seeded the session with:

<a b="c">d</a>

...and simply used that as the input for a vanilla copy of xmllint. I was merely hoping to stress-test the very basic aspects of the parser, without getting into any higher-order features of the language. Yet, after two days on a single machine, I found this buried in test case #4641 in the output directory:

...<![<CDATA[C%Ada b="c":]]]>...

What the heck?!

As most of you probably know, CDATA is a special, differently parsed section within XML, separated from everything else by fairly complex syntax - a nine-character sequence of bytes that can't be realistically discovered by just randomly flipping bits.

The finding is actually not magic; there are two possible explanations:

• As a recent "well, it's cheap, so let's see what happens" optimization, AFL automatically sets -O3 -funroll-loops when calling the compiler for instrumented binaries, and some of the shorter fixed-string comparisons will be actually just expanded inline. For example, if the stars align just right, strcmp(buf, "foo") may be unrolled to:
  
  

  
  cmpb   $0x66,0x200c32(%rip)        # 'f'
  jne    4004b6 
  cmpb   $0x6f,0x200c2a(%rip)        # 'o'
  jne    4004b6 
  cmpb   $0x6f,0x200c22(%rip)        # 'o'
  jne    4004b6 
  cmpb   $0x0,0x200c1a(%rip)         # NUL
  jne    4004b6 
  

  
  
  ...which, by the virtue of having a series of explicit and distinct branch points, can be readily instrumented on a per-character basis by afl-fuzz.
  
  
  
  
• If that fails, it just so happens that some of the string comparisons in libxml2 in parser.c are done using a bunch of macros that will compile to similarly-structured code (as spotted by Ben Hawkes). This is presumably done so that the compiler can optimize this into a tree-style parser - whereas a linear sequence of strcmp() calls would lead to repeated and unnecessary comparisons of the already-examined chars.
  
  
  (Although done by hand in this particular case, the pattern is fairly common for automatically generated parsers of all sorts.)
  

The progression of test cases seems to support both of these possibilities:


<![

<![C b="c">

<![CDb m="c">

<![CDAĹĹ@

<![CDAT<!

...


I find this result a bit spooky because it's an example of the fuzzer defiantly and secretly working around one of its intentional and explicit design limitations - and definitely not something I was aiming for =)

Of course, treat this first and foremost as a novelty; there are many other circumstances where similar types of highly verbose text-based syntax would not be discoverable to afl-fuzz - or where, even if the syntax could be discovered through some special-cased shims, it would be a waste of CPU time to do it with afl-fuzz, rather than a simple syntax-aware, template-based tool.

(Coming up with an API to make template-based generators pluggable into AFL may be a good plan.)

By the way, here are some other gems from the randomly generated test cases:

<!DOCTY.

<?xml version="2.666666666666666666667666666">

<?xml standalone?>

Read More

Bitcoiners: Surely we can do Buiter than this?

Willem Buiter has a very nice piece critiquing the Swiss Gold Initiative; see here.

Unfortunately, Buiter starts talking about Bitcoin, making false analogies between the cryptocurrency and gold. He should have just focused on gold.

As it turns out, both gold and Bitcoin do share some important characteristics. I've written about this here: Why Gold and Bitcoin Make Lousy Money.

The false analogy is in equating the mining of gold with the mining of bitcoin. Paul Krugman made the same mistake here: Adam Smith Hates Bitcoin. Here is the offending passage in Buiter's notes:

John Maynard Keynes once described the Gold Standard as a “barbarous relic”. From a social perspective, gold held by central banks as part of their foreign exchange reserves merits the same label, in our view. The same holds for gold held idle in private vaults as a store of value. The cost and waste involved in getting the gold out of the ground only to but it back under ground in secure vaults is considerable. Mining the ore is environmentally damaging, especially if it involves open pit mining. Refining the gold causes further environmental risks. Historically, gold was extracted from its ores by using mercury, a toxic heavy metal, much of which was released into the atmosphere. Today, cyanide is used instead. While cyanide, another toxic substance, is broken down in the environment, cyanide spills (which occur regularly) can wipe out life in the affected bodies of water. Runoff from the mine or tailing piles can occur long after mining has ceased. 

Even though, from a social efficiency perspective, the mining of new gold and the costly storage of existing gold for investment purposes are wasteful activities, they may be individually rational. The same applies to Bitcoin. Its mining is socially wasteful and environmentally damaging.

No, no, no and no. This analogy is all wrong.

Let me be clear about this. Bitcoin costs zero to produce. If one had control over the protocol, one could instantly and costlessly create as many bitcoins as one wanted. No environmental waste, no effort needed. The same is not true of gold.

But wait a minute, you might say. Doesn't mining for bitcoins require effort, consume resources, etc.? The answer is, yes, it does. But this fact does not make the analogy correct (though one can certainly understand why the analogy seems to be correct). Let me explain.

The purpose of gold miners is to prospect for gold. The purpose of Bitcoin miners is not to prospect for bitcoins. The purpose of Bitcoin miners is to process payment requests. A bank teller also processes payment requests. To say that miners are mining for bitcoin is like saying that tellers are mining for dollars. Understand? Let me try again.

Gold miners prospect for gold. But they do not necessarily get paid in gold. In fact, if they work for gold companies, they are likely to get paid in dollars. But they could get paid in gold, or anything else, for that matter. How they get paid does not take away their basic function, which is to discover new gold.

Bitcoin miners, like bank tellers, process payments. Miners, like tellers, want to get paid for the service they provide. It really does not matter how they are paid. As it turns out, miners are paid in the form of newly-issued bitcoins (as well as old bitcoins offered as service fees by transactors). But this does not mean that they are "mining for bitcoin" any more than a bank teller is "mining for dollars."

But isn't mining for bitcoin "wasteful?" In a sense, yes, but again, the "waste" here is not the same as the waste associated with commodity money. Again, let me explain.

We live in a "second-best" world, where people lie and cheat. In a first-best world, money would not even be necessary (see my post here: Evil is the Root of All Money). It is unfortunate that we need Bitcoin miners (and tellers) to process payments. But the resources consumed in this process are necessary, given the safeguards that have to enforced to ensure the integrity of the payment system.

The waste associated with mining gold is that in principle, gold money can be replaced by paper money (and please, do not give some weird "out of thin air" argument; see here.) Paper money, like Bitcoin, and unlike gold, is (near) costless to produce.

Note: Of course, the limit on the supply of bitcoin is determined by a community consensus on following the protocol that adopts the 21M limit. Bitcoin advocates argue that this "hardwired" protocol that governs the supply of bitcoin is more reliable and less prone to political manipulation relative to existing central banking systems. This all may be true, but does not take away from my argument above concerning the false analogy between gold and bitcoin.

Read More