Wednesday, June 25, 2014

Detroit's Fake Crisis: Water is the Leash

I got my MS in Environment Studies from CU Boulder, and my thesis was basically about water utility management and regulation. The Detroit water shutoff is preposterous. Some basic research is all you need to see that they're targeting the poor rather than the wealthy, and that this is strategic in terms of moving union backers, workers, and voters out of city limits to make way for gentrification.

There are also rumblings of privatization, this being an effort to make the "public" water utility prove its ability to garner payment (which is what its bonds are rated on and its interest rates are determined by; see here for good coverage of the bond insurance companies and demands for their pound of flesh). And, if we know anything from experience, privatization will result in even more massive rate hikes than 8-9% and even less competent service for small account holders. And if insurance companies are suing to avoid a cramdown, at the end of the day, they're just upset about making a bad bet and having to pay for it.

But how bad of a bet is it? This is the 2012 "in words" audit of the 2012 financial statement of the Detroit Water and Sewage Department. The 2013 report is not available on their website.

It shows, basically, that as of 2012 the DWSD was doing fine.

Not great, given the depopulation of Detroit (as concerns about declining water sales show), but fine: its margins were 20% for water and 22% for sewer, more than enough to cover their debt obligations. Let's ask the report, shall we?
The current ratio (current assets divided current liabilities) of 1.9 for water and 2.21 for sewer (M) indicates that the system is in a good position to meet its short term obligations, a key measure of fiscal health.
And let us not neglect the figure which shows this in detail:

So, the "$5 billion in debt" the media keeps talking about is the COMBINED debt of the WATER ($2.6B) and SEWER ($3.3B) operations, which have different revenue streams... and besides, that's not a very big number. For comparison, the water utility that serves Las Vegas has $3.6 billion in debt, not counting any sewage operations (which is a separate utility). There are more people in Las Vegas now, true, but should we expect Detroit water bills to be the same as the nation's average per capita, given its recent depopulation and aging infrastructure? Of course not.

Any justification for cutting off users based on debt is bullshit. Public water utilities have debt for very large capital investments that they pay off for very long periods of time. To cover the interest, they deliver water to customers basically at cost (they're regulated, after all), plus a legally required margin to cover debts. Detroit followed the law, and in fact exceeded the required margin.

Now, if a utility has a revenue problem, there are many ways to deal with it. One is charging more of the people who can pay more (i.e. industrial, commercial, and large residential customers such as apartment complexes). You can also tier rates among classes of user so that larger users (e.g. in residential, white suburbanites with lawns) pay more for each marginal unit of water, thus subsidizing basic human needs (drinking, cooking, washing) for others. OH WAIT!!! DWSD does charge variable rates, but they charge LESS for each marginal unit you use, and their tiers are incredibly weak (there are only two, really). This is awful financial management, and represents the exact opposite of how you should organize your water rates if you have any sense of environmental and social responsibility.

Utilities can even vary charges for the kinds of connections and pipes demanded by users and property owners (as the Las Vegas utility did in the 2000s). DWSD does this, but obviously isn't charging enough of the people and corporations who can afford to pay. DWSD has not been shutting off commercial customers who owe them money. I wonder why.

There are many more ways to get people's attention than shutting off their water. Public utilities are unique in that they are public, and there can be assessed penalties in other areas of life, seizure of assets, liens on properties, for non-payment. What moral person denies anyone else the right to drink water? To literally survive? Has the power to give someone life, but instead takes it away?

I mean this deeply: after three days without drinking water a human being dies. I don't expect anyone to die of thirst, but to avoid it, people will leave their homes in short order if that home doesn't have water in it. Shut off the power, shut off the cable, shut off the heat even and people will gut it out. But if I can't flush, I'm probably not living in the house for long.

DWSD will not free up significant resources, or improve operating revenues, by cutting deliveries: the less water DWSD delivers, the less money they make (they worry about that out-loud in their reports!). They say a little more than half the accounts are paid? Then this shows how easily they could have gotten the money other ways. There is no conceivable way that cutting off water increases revenue for DWSD under their current model, since one less address receiving water is one less resident able to pay in the future (no water, no people). Asking people to pay now when they cannot only increases poverty. Cutting delivery (rather than assessing a fine or *reducing delivery* as a signal to pay up) only serves to depopulate Detroit further, thus destroying the city that demonstrated the greatest successes of Labor in America and inspired a generation of workers to organize.

This is about undermining unions: nothing more, nothing less.

Well, perhaps more (or less, depending...) Two other facts: Having your water cut off means child protective services can come and take your kid, and getting your water turned back on requires at least some of the following: deed to the property, lease agreement-notarized, mortgage documents, tax records, driver’s licence, social security cards, notarized statements from the owners of the property, background checks...

Water is the leash... tug it, and the people move.

The elites have the leash now, not the people (emergency financial manager/dictator, anyone?), and now the leash is getting shorter, rapidly.

So you tell me what's going on. Is it about debt? Debt that DWSD has plenty of ways to continue paying? Or is it about waging war on the poor, working, minority citizens of a great American city that showed the way towards workers' - and peoples' - rights?

Ask Romney. He'll tell you.

Thursday, June 12, 2014

Due Process in the Teaching Profession

The easy out is that this is clearly the legislature's purview and not a judge's, ruling is too broad. The discrimination grounds are not the result of tenure but the result of administrative districting and funding decisions. Tenure is really just legal protection and due process for teachers. They can and are fired. When bad teachers aren't fired it's usually because the administrator isn't doing their job either. When it boils down, the admins are complaining that they can't fire *cheaply*, not that they can't fire. The $50,000-$500,000 price is just the salary they would have to pay for one additional teacher anyway at $50K/yr to replace the one in review for the 1-10 years they say it can take. Is it unreasonable to pay the equivalent of an additional teacher to ferret out whether a given teacher deserves to be in the system for a year? I think so.

As I read it, the judge knew it would be appealed so didn't go too deep into it. He's doing a lot of hand waving it seems, calling arguments names rather than addressing them or doing the research to support the claims or being critical of the research presented. The claim that a bad teacher costs a student $1.4 million in earnings is absolutely insane, since most people never make that much in their lives.

At the end of the day, last-in, first-out is the most fair way to do layoffs precisely because of the evidence that the single most important factor in determining teacher quality is time on on the job. And as for the protracted firing process, the complaint is totally unreasonable when you consider the way teachers are paid over their career: nothing at first, then a lot at the end. They want you to stick around when you're really good. So if you're going to put someone into review, doing so for 1-5 semesters in a career of 60-80 seems perfectly reasonable to me. Retraining is cheaper and better in terms of community structure than looking for a new teacher in many cases.

As for time to tenure, this judge's arguments might seem reasonable on its face, but again, lets look at how those 4 semesters really go. An administrator (and the fellow teachers) can tell quite well whether a teacher is cut out for the profession after a year. If you make it about review in the 4th semester, then that first year is very important. If you don't have it, you're out. Most bad teachers fail in that time, most walk away. Those that still want to try, do, and usually do get fired and are told to try at another school or district, where they start their seniority clock (i.e. 4 semesters) over again. Should that be 6 semesters? Fine, whatever. Pay them more and they'll take it. As a bonus, you'll get better qualified people. PhD's will teach HS if you started pay at $60,000/yr. I know I would. Junior College wages. The irony in education is that we pay those who do the least important work the most and those that do the most important work the least. We might even have better outcomes if we paid professors what we pay pre-K and pre-K what we pay professors. But if you make it 6 semesters, you'll get more shitty teachers sticking around, clogging up the pipeline that will grow in breadth by an additional two semesters: how many more bad teachers would stick around for that third year? Would that increase their number overall? What's better policy in a state with so many people going through public teaching universities?

And let me say this... quality of teacher matters on the margin, but home environment matters more. If you have parents who value education, have resources, and don't fight in front of you, your teacher will matter little because the parents pick up the slack. If your parents don't value education, don't have resources, and/or fight all the time, then teachers will matter little because the kid isn't going to have the support. Poverty menaces minority students, not bad teachers. Teacher quality can boost child performance, but it does so from a baseline set in the home.

Basically, students could have sued and said that the state had failed to provide incentives for competent teachers and support for their "underperforming" schools. There is a long tail in universal public education. In fact, an argument could perhaps better be made that the state failed to place the most effective teachers in the right schools and has simply failed to invest in those schools; not that the state failed to fire certain teachers. If you fund those schools, admin doesn't have to spend as long or as much trying to fire people. This means you could clear the decks simply by funding the system as it is. It works.

Admin can do their jobs better if you fund the schools. Give them the tools they need, the resources, the bodies, to do the work.

And maybe the writing styles are different, but legal and academic writing seem to have different standards for evidence. My first year students give more careful consideration of counter arguments and the sources of evidence than this judge appears to have.

Of course, as always, any measure of teacher or student performance based on standardized tests is going to be wrought with all kinds of racial, gender, and class based biases. So let that be considered in our thinking on what a "failing" school is, or a "bad" teacher, when in reality these teachers are in many cases producing the best outcomes you could ask from any reasonable person. You try to manage 160 kids on free lunch each day, let alone teach them the history of a continent that systematically murdered their people... and then fractions and proper English. This throws into question the 9.54 month figure (can you really have that level of scientific precision? What does that imply in terms of significant figures? 9 Months, 16 Days, 17 Hours. Really?). I'll say it again: There is a long tail in universal public education. Charters don't have to take people with learning disabilities, and the ones in public schools often go undiagnosed. I haven't looked at the study itself, but it seems like someone is reaching a bit too far on that one. How many units behind are they? Speak in meaningful metrics, like what remedial summer work we would need to invest in to make up for this cost. Ultimately, this shows that whoever did this research is not thinking like a teacher, like an educator, but like a technocrat who is not asking a meaningful question for the classroom experience. Whoever came up with 9.54 months was looking for a headline, not something teachers could use.

Finally, the contention that 1-3% of teachers (workers) are ineffective is actually a testament to how well educational training in California is working. How many of those are first or second years who will be fired? How many will never get tenure? There will of course be an error rate, especially when the starting salary is a minimum $35,000K/yr, and rarely over $40,000/yr. Clearing that deck should simply be a cost of business. It's remarkable that they will go to such expensive lengths to avoid costs of $50,000k/yr to the system per teacher in review (and attack organized labor and through it working people on the whole who use unions as a backstop against overzealous capital and corrupt government). Or, following the math:

$50,000 per teacher per year x 2750 to 8250 teachers per year = $137,500,000 to $412,500,000 per year

There are about 10.8 million taxpayers in California. So paying a teacher's yearly salary to deal with the 1-3% of purportedly ineffective teachers (remember the flaws of testing) would cost $12.73 to $38.19 per taxpayer per year assuming this is a constant error rate in the production of teachers and not a build up (in which case this cost falls over time). Progressive and geographically targeted property tax structure, etc. makes this more effective social policy, and suddenly we're looking at a perfectly reasonable solution that doesn't require debasing working people and removing legal guarantees to due process for public educators.

Monday, April 21, 2014

CU Boulder Town Hall Campus Update

The University sends me the following:

My reply:

I have been at CU for five years now, working first on a terminal MS and now a PhD in Environmental Studies. I have been teaching in the classroom for four of those years, plus summers. I have five questions related to the overarching policies in place by CU.

1) What, if any, will be the University’s policy towards “student” athletes when/if they decide to organize? How will the University reconcile these policies with those towards graduate professionals, who do not have the right to organize but are explicitly employed as labor by the University?

2) I make 25% of what my students pay to take my classes. Where does the other 75% go? How is it enhancing their education/improving the “product” you sell? Is that other 75% providing three times the educational value that I provide in my classroom? If not, why is it being spent?

3) What, if anything, will the University do to address its inability to treat criminal sexual assault and harassment as a crime, not a “code violation”? Will the University reform the Office of Student Conduct, the Office of Discrimination and Harassment, and CUPD to create a true criminal justice system for crime on this campus, or will the University continue to shelter rapists, violent sexual offenders, and verbally abusive people behind a kangaroo court-like “investigation” where victims' rights are not respected and “restorative justice” programs allow the wealthy to write papers for things the poor go to jail for? When will the University take seriously its mission to educate the whole person, and teach personal responsibility using the same consequences one would face in the real world?

4) We already see the University defending its hateful “visiting scholar of conservative thought” through its silence after his disgusting diatribes against people who are not like him on this campus. What is the University doing to ensure that the recent Regent’s policy surrounding political ideology and affiliation is not used to silence dissent on campus or allow for hate speech? How is this protected behavior? If I said anything remotely like what Steven said I would be fired immediately; why hasn’t the University fired Steven Hayward, and what will the university do in the future when an explicitly conservative political puppet harasses and demeans the student population of CU?

5) When will the University divest from fossil fuels and end its support of (a) the international criminal gangs/hit squads paid off by fossil fuel companies to murder environmentalists and civil rights leaders, (b) policies here in the US that are anti-worker and counter to the public good, and (c) organized crimes against humanity through the complete and utter destruction of the ecosystem upon which out species depends?

I don’t expect you to answer these questions because they’re not the kind that build CU’s brand. But an honest assessment of them will be crucial in determining the future of CU’s “climate” for undergraduates and graduate professionals, which you are apparently concerned about now that a Republican regent has released the hounds.

With a disposition not meant to manipulate you emotionally,


Friday, March 7, 2014

Projected Climate Change and Impacts in the Western US: Pielke Jr. v. Podesta

In light of recent events, it seems appropriate for me to provide the following, which is an answer I gave on my preliminary exam...

A) What are the most robust findings for how climate might change in the future in the western US according to the IPCC?  Discuss sources of uncertainty in determining climate change and associated impacts that are relevant to water resources;

Projected Climate Change and Impacts in the Western US

The Western United States is an arid region (Woodhouse, 2010) (Frisvold in Colby, 2011), receiving insufficient precipitation to raise crops without irrigation on the vast majority of the land. Furthermore, its climate is historically highly variable across space and time (Seager, 2009) (MacDonald, 2005). The region does have significant groundwater resources, however they have been strained over the past century and the resource is becoming more expensive and environmentally damaging to access (Ashley, 1999) (Schlager et al., 2006). Finally, ecosystem primary productivity in the Western US is dominated by forests, which has implications for a changing climate especially related to fire and runoff patterns (Schwinning, 2008) (Williams, 2010).

The most robust findings for how climate might change in the future in the western US according to the IPCC are largely in line with the present climate regime: climate is projected to remain highly seasonal (i.e. wet, cold winters and hot, dry summers), but with higher temperatures and less streamflow (IPCC, 2007) (IPCC, 2013). The table below summarizes the impacts over Western North America (roughly everything West of Utah) or North America (US and Canada), unless otherwise indicated.

Table 1. Climate change and its impacts (except sea-level rise) on the Western US, taken from IPCC (2007, 2013).

Experienced or Anticipated Changes
Experienced or Anticipated Impacts
By 2046–2065, warming in all regions exceeds the natural variability estimate for all models. Thus it is very likely the warming signal will be large compared to natural variability in all NA regions throughout the year by mid-century.  (IPCC, 2013:14-41) For most combinations of model, scenario, season and region, warming in the 2010 to 2039 time slice will be in the range of 1 to 3oC. (IPCC, 2007:626) Late in the century, projected annual warming is likely to be 2 to 3°C across the western, southern, and eastern continental edges, but more than 5oC at high latitudes (Christensen et al., 2007: Section The projected warming is greatest in winter at high latitudes and greatest in the summer in the south- west U.S. (IPCC, 2007:627)
Projected warming in the western mountains by the mid-21st century is very likely to cause large decreases in snowpack, earlier snow melt, more winter rain events, increased peak winter flows and flooding, and reduced summer flows (IPCC, 2007:62). Warming is very likely to place additional stress on groundwater availability, compounding the effects of higher demand from economic development and population growth (medium confidence) (IPCC, 2007:55)
There is generally less confidence in projections of precipitation than of temperature (IPCC, 2013:14-38)
A robust evaporation increase characteristic of mid-latitude continental warming (Seager et al., 2007; Seager and Vecchi, 2010) … the CMIP5 models still show a strong decrease in soil moisture here (Dai, 2013), due to increasing evaporation. (IPCC, 2013:14-42)
Rising temperatures will diminish snowpack and increase evaporation, affecting seasonal availability of water. (IPCC, 2007:55)

The fraction of annual precipitation falling as rain (rather than snow) increased at 74% of the weather stations studied in the western mountains of the U.S. from 1949 to 2004 (Knowles et al., 2006). (IPCC, 2007:622)

Winter precipitation increases extend southward into the USA (northern portions of SREX regions 3–5; Neelin et al., 2013) but with decreasing strength relative to natural variability. (IPCC, 2013:14-42)
This behaviour is qualitatively reproduced in higher resolution simulations (Figure 14.18). (IPCC, 2013:14-41)
When downscaled, CMIP3 models showed less drying in the region (Gao et al., 2012c) and an extreme precipitation increase (IPCC, 2013:14-42) CMIP5 models do not consistently show such a precipitation decrease in this region (Neelin et al., 2013). This is one of the few emerging differences between the two ensembles in climate projections over NA. (IPCC, 2013:14:42)
Streamflow over the last century has decreased by about 2%/decade in the central Rocky Mountain region (Rood et al., 2005). IPCC, 2007:621)
Climate change will constrain North America’s over-allocated water resources, increasing competition among agricultural, municipal, industrial and ecological uses (very high confidence). (IPCC, 2007:55) Major challenges are projected for crops that are near the warm end of their suitable range or which depend on highly utilised water resources. (IPCC, 2007:15)

An increase in drought frequency (Sheffield and Wood, 2008; Gutzler and Robbins, 2011) (IPCC, 2013:14-42)
Climate change will constrain North America’s over-allocated water resources, increasing competition among agricultural, municipal, industrial and ecological uses (very high confidence). (IPCC, 2007:55)

Spring and summer snow cover has decreased in the U.S. west (Groisman et al., 2004). April 1 snow water equivalent (SWE) has declined 15 to 30% since 1950 in the western mountains of North America, particularly at lower elevations and primarily due to warming rather than changes in precipitation (Figure 14.1a) (see Mote et al., 2003; Mote et al., 2005; Lemke et al., 2007: Section (IPCC, 2007622)

Warming, and changes in the form, timing and amount of precipitation, will very likely lead to earlier melting and significant reductions in snowpack in the western mountains by the middle of the 21st century (high confidence) (Loukas et al., 2002; Leung and Qian, 2003; Miller et al., 2003; Mote et al., 2003; Hayhoe et al., 2004). (IPCC, 2007:627)

Streamflow peaks in the snowmelt dominated western mountains of the U.S. occurred 1 to 4 weeks earlier in 2002 than in 1948 (Stewart et al., 2005).  (IPCC, 2007:622)
In projections for mountain snowmelt-dominated watersheds, snowmelt runoff advances, winter and early spring flows increase (raising flooding potential), and summer flows decrease substantially (Kim et al., 2002; Loukas et al., 2002; Snyder et al., 2002; Leung and Qian, 2003; Miller et al., 2003; Mote et al., 2003; Christensen et al., 2004; Merritt et al., 2005).
(IPCC, 2007:627)

Extreme Heat
Warming generally leads to a 2- to 4-fold increase in simulated heat wave frequency over the 21st century (e.g., Lau and Nath, 2012). (14-41)  Warm extremes across North America are projected to become both more frequent and longer (Christensen et al., 2007: Section (IPCC, 2007:627)
Cities that currently experience heatwaves are expected to be further challenged by an increased number, intensity and duration of heatwaves during the course of the century, with potential for adverse health impacts. (IPCC, 2007:15)

Warmer summer temperatures are expected to extend the annual window of high fire ignition risk by 10-30% (IPCC, 2007:619)
Disturbances such as wildfire and insect outbreaks are increasing and are likely to intensify in a warmer future with drier soils and longer growing seasons (very high confidence) (IPCC, 2007:619)

It is very likely that ENSO will remain the dominant mode of interannual variability with global influences in the 21st century, and due to changes in moisture availability ENSO-induced rainfall variability on regional scales will intensify. (IPCC, 2013:14-23)
There is medium confidence that ENSO-induced teleconnection patterns will shift eastward over the North Pacific and North America. There is low confidence in changes in the intensity and spatial pattern of El NiƱo in a warmer climate. (IPCC, 2013:14-23)
ENSO shows considerable inter-decadal modulations in amplitude and spatial pattern within the instrumental record. Models without changes in external forcing display similar modulations, and there is little consensus on whether the observed changes in ENSO are due to external forcing or natural variability (IPCC, 2013:14-23)

Sources of Uncertainty in Climate Change Predictions

Data Collection Uncertainties

The IPCC (2013) identifies uncertainties related to the raw observational data itself:

Measurements have changed in nature as demands on the data, observing practices and technologies have evolved. These changes almost always alter the characteristics of observational records, changing their mean, their variability or both, such that it is necessary to process the raw measurements before they can be considered useful for assessing the true climate evolution… the vast majority of the raw observations used to monitor the state of the climate contain residual non-climatic influences. Removal of these influences cannot be done definitively and neither can the uncertainties be unambiguously assessed. Therefore, care is required in interpreting both data products and their stated uncertainty estimates (2-8).

The approaches taken to reduce this uncertainty are, “redundancy in efforts to create products; dataset heritage; and cross-comparisons of variables that would be expected to co-vary for physical reasons, such as land surface temperatures and sea surface temperatures around coastlines” (2-9). As a general matter, removing noise to identify signal is standard practice in data analysis. Of all the complicating factors, this would seem the least difficult to address. That said, what error there is will hopefully be non-systemic and small; errors into these data would have significant effects on the outcomes of parameterization efforts. Page 9-61 also notes that “In some cases, insufficient length or quality of observational data makes model evaluation challenging, and is a frequent problem in the evaluation of simulated variability or trends.”

Data Handling Uncertainties

The IPCC (2013) does not address this in great detail, but it does discuss “dataset heritage” (above), as well as make this point:

The uncertainty in observational records encompasses instrumental / recording errors, effects of representation (e.g., exposure, observing frequency or timing), as well as effects due to physical changes in the instrumentation (such as station relocations or new satellites). All further processing steps (transmission, storage, gridding, interpolating, averaging) also have their own particular uncertainties. Since there is no unique, unambiguous, way to identify and account for non-climatic artefacts in the vast majority of records, there must be a degree of uncertainty as to how the climate system has changed (2-8).

Given that, we should clearly be aware of how data has been treated before it is used in modeling or generating historical trends. This should help prevent “error propagation” (9-60).

Data Processing Uncertainties

According to the IPCC (2013), “Some model errors can be traced to uncertainty in representation of processes (parameterisations). Some of these are long-standing issues in climate modelling, reflecting our limited, though gradually increasing, understanding of very complex processes and the inherent challenges in mathematically representing them” (9-60). That said, they do have increased confidence in the newer Atmosphere Ocean Global Climate Models (IPCC, 2013:9-3). However, models might miss key thresholds and generalize across vast areas (IPCC 2013). Furthermore, “Future anthropogenic emissions of greenhouse gases, aerosol particles and other forcing agents such as land use change are dependent on socio-economic factors including global geopolitical agreements to control those emissions,” (12-10) meaning there are uncertainties in human side.

Other issues are that “Regional climate projections are generally more uncertain than projections of global mean temperature” (14-38) and bio-topography may override the climate signal (“sensitivity to resolution” (9-60)). In addition, clouds are not well understood: The simulation of clouds in climate models remains challenging. There is very high confidence that uncertainties in cloud processes explain much of the spread in modelled climate sensitivity. Nevertheless, biases in cloud simulation lead to regional errors on cloud radiative effect of several tens of watts per square meter” (9-3). In sum, “Projections of climate change are uncertain, firstly because they are primarily dependent on scenarios of future anthropogenic and natural forcings that are uncertain, secondly because of incomplete understanding and imprecise models of the climate system and finally because of the existence of internal climate variability” (12-8).

Output Application Uncertainties

Impacts and implications are uncertain because the climate system may change and cause totally novel regimes. Furthermore, ecosystems themselves are not stable over time and are highly unpredictable. There are feedbacks between climate and ecosystems that exhibit properties not easily predicted. For example, Cody (2011) identifies that, “melting snowpack up-regulated by temperature and dust deposition increases the vulnerability of forests to summer fires. That itself increases vulnerability to bark beetles. And beetles, by creating stands of dead stands of trees, increase fire vulnerability” (21). This kind of positive feedback may not always be accounted for in climate models, but landscape level outbreaks could have significant impacts on the water cycle through impacts on forest cover.

Human systems are highly variable and unpredictable as well. For example, if controlled burns were implemented in the situation above, it may exacerbate exposure to bark beetle, and thus doom the remaining trees. Predicting this intervention in a climate model would be impossible. This example may seem superficial, but consider the vast irrigation works of the Great Plains and American West. How are we to predict the future water usage, and thus impacts of climate change? What of groundwater overdraft, or oil shale development? This issues illustrate the point that impacts much removed from changes to temperature and precipitation regimes must be interpreted with added caution.