I wrote >>What is the problem with using some (but not all) government statistics
as a half-bad/half good way of understanding what's going on, in conjunction with other information and reasoning?<<
Sabri Oncu wrote, > >Life is not as rational as you think it is.<<
Who said that life was "rational"? One of things we should strive for is for life to be more rational.
(I'll beg off on the definition of "rational" for now -- who has the time? Put it this way, it's not the instrumental rationality of "Western enlightenment" thinking. There's more than one kind of rationality.)
Tom Walker:
> For that matter, the rate's rationality may not be all its cracked up to be.
> My answer to Jim's question is: nothing is wrong if we fully acknowledge the
> limitations of the government statistics -- or any statistics -- to
> "measure" the phenomena they purport to measure. The problem
> is that we _do not_ acknowledge those limits but become indignant or
> uncomprehending when someone once again raises the usual objections, let alone
> unusual ones.
I _do_ acknowledge these limits, as does Doug (in my experience). Who is this "we" you refer to? I really hate being a straw man.
Using statistics intelligently (or "scientifically") always involves two different things: (1) actually using them and (2) being aware of the limitations of the statistics. This is a key point that critics of Doug and myself on this issue miss.
> The basis of [one version of] rationality is non-contradiction: "the same
> person cannot at the same time hold the same to be and not to be." By the same token,
> presumably, the same person cannot be employed and unemployed
> at the same time. Voila, we have a statistic!
This ignores the fact that there are statistics on "involuntary part-time workers," who can be seen as both unemployed and employed at the same time.
> However the same person *can* be employed and unemployed at successive
> moments. The definition of unemployed includes that the person is actively
> looking for work and therefore, implicitly at least, will be employed at
> some time in the future.
If you look, you can find some stats on people's experience over time with unemployment, and I presume, employment. You can also find estimates of hours worked per week, too.
> To qualify for unemployment benefits, one must have
> worked a minimum number of weeks in the recent past.
In the U.S., at least, there is no connection between such eligibility and officially being counted as "unemployed."
> Thus unemployment is only unemployment in relation to a past and/or future
> employment, usually both but not certainly either. In other words, "the state of
> unemployment" implies a movement toward or away from itself.
this last sentence doesn't make any sense. But it's quite easy to get a time series of unemployment data (measured in different ways). In fact, the time series makes more sense, as long as one doesn't focus on month-to-month changes: a year-to-year increase in the official unemployment rate has a very simple meaning: all else constant, workers are being screwed. Of course, all else isn't always constant, so that workers can be screwed without unemployment rates rising.
> Dynamically, the concept relies on contradiction. Only statically does it
> appear to be non-contradictory. The statistic necessarily treats
> unemployment at rest, so to speak. A statistic gives a static
> picture. It is no coincidence that both words begin with the same four letters.
how about "statistics"? if you take a time series of statistics, it doesn't provide a static picture, even though it it "begin[s] with the same four letters" as that word.
The monthly unemployment rate does represent a snap-shot. But put enough of them together, you get a movie, or at least a slide-show.
> Zeno's paradox shows the problems inherent in treating a
> moving object as if it occupies successive positions of rest. I won't go into the details.
> Contradiction isn't necessarily a bad thing, it simply points to the limits
> about what we can say about dynamic phenomena.
the contradiction disappears if you realize that changes in unemployment rates are more important than the level.
> The illusion of a dynamic picture of unemployment is created by placing last
> month's or last year's static picture beside this month's. We say
> "unemployment is up" or "unemployment is down" when we really
> have no idea of how many employed people are moving toward unemployment,
> and how fast they are moving in that direction or conversely how many
> unemployed people are moving how rapidly toward employment.
we don't really know "many employed people are moving toward unemployment, and how fast they are moving in that direction or conversely how many unemployed people are moving how rapidly toward employment" from the unemployment rate, but that doesn't mean we can't find out -- or at least get some idea -- from other statistics that are available. Absolutely no-one that I know of who thinks that one can glean useful information from the BLS labor-market statistics thinks that the (official) unemployment rate is the _only_ number that's worthy of our attention.
> I'll just mention in passing that gross movement into and out of the labour
> force typically swamps net change in the ratio between employed and
> unemployed labour force participants. In fact, people in the
> U.K. who have studied this have found that much of the movement occurs directly from
> non-participation to employment or from employment to
> non-participation and not incrementally between non-participation, unemployment and
> employment.
so much is obvious. I think you should look at Dean Baker's monthly commentary on the BLS press release on unemployment. It shows us how to interpret these numbers.
> The U rate thus refers to something quite different than what is happening.
> (The expected response here is that "we know this" but it is
> useful as an "indicator" of what is happening. The caveats on an indicator
> have worn smooth, plus or minus 3%, 19 times out of 20, before that
> indicator enters into general circulation.)
we don't have to wear the caveats down. Instead, an intelligent observe has to emphasize them.
> Also according to the principle of non-contradiction, a person cannot be an
> unemployed certified aircraft mechanic at the same time he or
> she is an employed telephone salesperson, for as little as one hour a
> week. Perhaps Jim or Doug would like to point out that we can tease out the
> extent of underemployment or discouragement from various supplementary
> sources. Indeed we can tease out, somewhat, the extent of these but not their
> intensity.
it's very hard to measure "underemployment," especially since its definition is more "contested" than most. But what I've seen is that it tends to rise with the various measures of unemployment, at least in the U.S.
> Subjectively, it is the intensity of unemployment or underemployment that
> matters (e.g., "did I make enough this month to pay the rent") and here you
> have a phenomenon that is utterly absent from the numbers.
that's why we don't focus on a single number.
> Don't ask me what data would describe this intensity of un/underemployment.
> It is a qualitative fact and not a quantitative one. One might say, given
> the bounds of rationality, that the government statistics are
> not all that far from "the best we can do" quantitatively, especially if
> we are hoping for a single number that summarizes the whole damn thing.
no one that I know of has that hope.
> Admittedly 5.6% tells me a whole lot more than some number pulled out of the
> air, say 1068 or six of one, half a dozen of the other. A large part of what that 5.6%
> means to me, though, is constituted by what I know the number
> doesn't tell me. Namely, it doesn't tell me that unemployment is "down"
> this month (or up this month). Unfortunately that is *precisely* how it is
> talked about in the media, by government officials etc. and thus that is the
> discursive frame imposed on it.
are there any government officials or representatives of the mainstream media on pen-l? if not, go talk to those people about these matters. I'm quite familiar with them, so I don't need to be told.
> Remember the [Walkerian-Aristotelian] definition of rationality: not believing
> something to be and not to be at the same time. If the discourse about
> unemployment rate were rational, it wouldn't be about ups and downs.
I don't understand this at all.
> Even when we are talking about the measurable equivocations of the U rate,
> we may be implying 5 to 40 pages of inaccessible technical discussion,
> ambiguous interpretation and scholarly disputation. All that jazz has
> nothing on the deliciously clear and unambiguous nobility of a single
> number. Who really cares how many unemployed angels can dance
> on the head of pin?
Sophism lives.
> Presumably the discourse about unemployment and an unemployment rate started
> out as a criticism of conditions and a rationale that "somebody ought to do
> something". There is thus a performance evaluation component
> to it and a reformist agenda underlying it. The problem with using an
> index number or test score for performance evaluation is that people can
> start teaching to the test or cream-skimming.
hey, if the gov't actually thought that a low unemployment rate was a good thing, it might be an improvement! Of course, there are ways to lower the rate, such as putting people in prison camps, that aren't especially attractive. But I doubt that governments do such things because they target a low unemployment rate as a goal.
In any event, this just tells us something that intelligent observers of BLS statistics already knew: no single statistic can sum up the entire labor-market experience.
> Getting the numbers right becomes
> the reform agenda -- the bottom line, as it were. The manipulation
> involved at the policy level aimed at getting the numbers right can very
> appropriately be thought of as Enronesque. No laws need be broken. No
> incorrect numbers need be entered in the books. Only a corruption of purpose and
> meaning need take place.
can you name someone who wants to focus entirely on the U rate?
> In short, it is possible to produce a "lower" unemployment rate while
> making employment more precarious and less rewarding and unemployment more
> personally and financially painful. This is after all what the right-wing
> think tanks and the OECD jobs strategies have in mind when they talk about
> 'flexibility'. Inflexible are things like layoff protections, generous
> insurance schemes and union membership. Flexible are things like
> just-in-time production and contingent employment.
this makes sense: Mexico, for example, has very little open unemployment (as measured by BLS standards) because it has very little in the way of unemployment insurance, so that we could drive U toward zero by moving toward a Mexico-style system.
But no-one that I know of says we should ignore the social context of unemployment measures. Are you saying that because we cite official statistics, Doug and I are advocates of neo-liberalism?
> Did I say "indicator" several paragraphs above? Indicator of what? Is the
> unemployment rate an indicator of what some other, more comprehensive set of
> statistics might tell us if we had the time and energy to collect and
> analyize them? Or are we using the unemployment rate as an indicator of the
> precariousness of employment and the misery of unemployment?
> But wait. If the judicious implementation of precariousness and misery can
> lower the rate of unemployment, is it rational to believe that a fall in the
> unemployment rate is an indicator of a decline in precariousness and misery.
this tells us we should ignore rising measures of unemployment produced by the BLS?
> Daniel Davies said about as much in many fewer words.
brevity is the soul of wit.
JD
