Wednesday, October 8, 2014

What problem does the postdoc pose?

A couple of days ago, a dialog began on an academic listserve.  An article in the Boston Globe citing a glut of postdocs serving a crisis served as the impetus.  (See the National Postdoctoral Association for a definition of postdoc)

The article addresses the"plight" of the biomedical postdoc,
The plight of postdocs has become a point of national discussion among senior scientists, as their struggles have come to be seen as symptoms of broader problems plaguing biomedical research. After years of rapid growth, federal funding abruptly leveled off and even contracted over the last decade, leaving a glut of postdocs vying for a limited number of faculty jobs. Paradoxically, as they’ve gotten stuck, the pursuit of research breakthroughs has also become reliant on them as a cheap source of labor for senior scientists.   ...
Their progress is very poorly tracked; the leader of a national report on the state of postdocs has called them “invisible people.” The National Institutes of Health estimates there are somewhere between 37,000 and 68,000 postdocs in the country. Salaries vary, but rarely reflect their level of education. The NIH stipend ranges from $42,000 a year for a starting postdoc, up to $55,272 for a seventh year.
The listserve discussion introduced a need to lobby Congress on behalf of postdocs and the assertion that postdocs experience very real suffering and systematic oppression.

I am not a biomedical postdoc.  I am a postdoc in the social sciences so maybe, I just don't know how it is.  But, my personal feeling is that identifying the nation's doctoral graduates in a highly technical field as a suffering, plighted group is less than desirable, demonstrates an odd view of the world and need for prioritizing problems.

Based on the Globe report, postdocs appear to be receiving a stipend that outside of the field could be considered a salary.  While benefits may be questionable (I just don't know) and reliability of work unstable, well, that seems a very real problem facing Americans generally, not just postdocs.  Finally, we don't all get to be rockstars, super models, trust-funders or socialites... just like we don't all get our own labs.

But what a great nation is this! Society allots funds to support the advanced training of researchers so they may pursue their dream career that has potential to eventually demonstrate societal benefit.  

My brief involvement in the listserve conversation expressed my feelings of pshaw for a designation of postdoc suffering.  For this I have some regrets.  I realize it is not my place to judge the worthiness of felt plight.

And what a great nation is this!  Those that feel wronged are justified in their feelings and to the extent by which they believe it merits Congressional attention they have the freedom to make their opinions heard.  

So, I, professionally, offer a different discussion:

Career dissatisfaction of University graduates seems a going concern.  This extends even to those with the highest degree attainments, PhDs.  All face a difficult job market, income inequality, lower than hoped for wages, etc.

The debate about PhD and postdoc "glut" has taken place, at least, since the early 1980's.  For instance, William Zumeta of the University of Washington, published a Higher Education article in 1982 titled, "Doctoral Programs and the labor market, or how should we respond to the 'PhD glut?"  At that time, Zumeta argued,
at the margin, some important benefits of doctoral programs have generally been understated.  At the same time marginal costs with respect to doctoral enrollments (and thus potential savings from enrollment reductions) have tended to be exaggerated by higher education policymakers in recent years.
A couple of years later, Zumeta published an article that saught to better understand not just the quality of the postdoc experience, but the quality of the postdoc.  The widespread belief at the time was the best doctoral graduates went right on to faculty positions.  Lesser graduates went on to postdoc.  But if the scientific engine depended on the postdoc worker than what did this all say for the quality of science?

While Zumeta called for further data collection and analysis he also stated, "cause for concern" existed.

Research in this field has continued and there is today, a good amount of information on graduate job attainment, experience, etc.

Perhaps the production of PhDs is a subsidized product, like corn.  The public highly values the idea of the farmer and farming and highly values food production.  So, the public seeks to provide some stability to the corn market.

Similarly, the public highly values education, research and the production of science.  So, society provides some stability to the higher education market.  We can argue about how desirable the situation is in corn and PhDs.

Are we using postdocs as a proxy for arguing about funding for research generally and biomedical specifically?  Probably.  

I think the whole discussion may make a bit more sense in the historical context of science policy.

It is after all, the societal goals for science that has fostered the growth of today's research army.  And to the extent the institution of science has faltered on maintaining its own integrity, then perhaps this has led to unrealistic expectations of the institution and/or a general poor quality experience for the postdoc.  

But, I'm not really sure that PhDs interested in research are so different from aspiring actors.  The difference being the researcher is not solely interested in fame (most aren't), but in the production of work that aids society in bettering itself creatively, pragmatically and respectably.  Some of us get there.  Others of us don't.  Much like Jennifer Aniston and that other actor you've never heard of.

I feel grateful that society supports research in the quest for societal benefit and and values broad access to higher education.  Still, it is highly likely that the whole system could work far better.

Monday, October 6, 2014

My Average Observation in the Insurance Journal


Recently, I've been trying to catch up on some readings in the classics.  I love revisiting what was said in the past and consider it in the context of the present and what is said about the future.  It's sort of remarkable how much things remain the same or how today is a caricature of past foresights.

For instance, was George Orwell really that far off?  I'd say the Jetsons characterize today pretty well too...  

In any case, one of these classics, The once and future school of public policy, was by the legendary political scientists, Aaron Wildavsky.  I know I read this for class while in graduate school, but I've found that things take on new depth of meaning after school and one has the patience to more fully consider the words people use.

Amongst, Wildavsky's thoughtful comments, he pointed out the following,
The best bet always is that the future will be like the past plus or minus 5 percent.
This was not the main lesson of the article, but nonetheless.

Today, I have an opinion in the Insurance Journal that falls in line with Wildavsky's observation of life for the most part.

Based on the data provided in charts by Dr. Hartwig of the III, I argue that by adding some context it would appear that future insurance profitability peaks will be, like the past, about average.

Friday, September 26, 2014

City/ Not City



Previously, I briefly discussed Miami-Dade's Master Plan and what looked to me like double speak.  In particular, I concerned myself with the concept of urban expansion versus that of sprawl.  A brief look at the academic literature demonstrated that defining sprawl is difficult and no one accepted metric exists.  So, for Miami-Dade to determine they have urban expansion rather than sprawl is futile.  

The video above demonstrates that defining a city is no simple task, either.  The researcher in the video indicate that cities tend to share similar characteristics but on different scales.  Still the clear identification of city/not city is elusive.    

Thursday, September 25, 2014

Guest Post at Artemis


I have a guest post on the Artemis blog.  The blog and its associated website is a really amazing source for information on insurance linked securities, particularly cat bonds.  I've found it useful in my research and improving my understanding of the risk sharing and trading enterprise.  

The piece is about model risk. Despite grand efforts to help risk managers feel better about the uncertainty in their models, the risk that model is just wrong remains persistent.
In response to client demand, catastrophe modelers are offering improved access to model components and ease of model blending, morphing, fusing, etc. Most notable of efforts are those of RMS, Karen Clark and Company (KCC) and Lloyd’s. RMS(one) promises to provide users with access to over 300 probabilistic models, whereas KCC’s RiskInsight enables users access to internal assumptions. Lloyd’s Oasis offers users choice in “a set of plug-and-play components.”

These efforts are aimed at resolving concerns about model risk but do not actually help to reduce or control model risk. Improved ability to manipulate vendors’ models may buffer companies from volatility produced by model updates. But that volatility is produced by changes in the decision making by the model vendors and their judgments about how best to create a model.

The ability to create one’s own theory on how best to estimate a given risk does not make that theory an accurate representation of reality. 
The rest of the piece is here.

Tuesday, September 23, 2014

Consulting Services for FCHLPM: Sunshiny or Convoluted?

http://borgman.cincinnati.com/
The state of modern science is such that one can legitimately piece it together in a myriad of ways to develop one's own unique perspective on a given issue.

This provides risk model vendors with with opportunity to develop their own secret sauce.  It  provides the vendor with their slice of the market and is therefore, fiercely protected.

Florida has "sunshine laws" which mandates that much of the government's going-ons, particularly meetings, becomes public record.   Florida's Sunshine Laws are in accordance with the Government in the Sunshine Act (1976) signed into law by President Ford.  It was introduced to the Senate- just after Nixon left office- by Florida's very own Lawton Chiles.  Presumably due to Nixon's legendary meeting secretiveness, the bill passed fairly quickly.

The Sunshine Act compliments the 10 years senior Freedom of Information Act (1966) which makes a bunch of government information public with some exceptions.  One exception is company trade secrets.  

So, there is a difficulty: Company's want to keep their information secret, FOIA allows them to, but once they have a "meeting" with public officials it becomes public information.  In order to resolve this difficulty in regards to catastrophe model vendors' secret science, the FCHLPM (those responsible for reviewing catastrophe model science) hold their meetings off public property at the vendors' office.

(Later today: It was brought to my attention that I may not be entirely clear on how FOIA and the Sunshine Law actually work.  However, I am still under the impression that the reason for site meetings [at the vendors' offices] is to avoid public disclosure of trade secrets.  As to how this legally works out, I'm not sure.  Some particulars on how and when FCHLPM things are exempt from public disclosure is here under section g.)   

However, models are becoming increasingly complex.  They have many more moving parts today then when first adopted by the industry in the early 1990s.   For example, under FCHLPM 2000 Standards, AIR submitted a report of 306 pages and under 2011 Standards the report was 404 pages.  

It takes a long time to review this amount of information perhaps longer than a site visit allows.  Besides, FCHLPM members have full time day jobs in addition to Commission responsibilities.  

To resolve this time consuming complexity problem (or so the story goes), the SBA has recently announced a request for consulting services
The SBA is seeking approximately fourteen (14) to sixteen (16) different vendors in seven areas of discipline (Meteorology, Structural Engineering, Coastal Engineering, Hydrology, Actuarial Science, Statistics, and Computer Science) to provide consulting services to the Commission. Respondents should be prepared to provide, at a minimum, the services described in the Report of Activities
The SBA financially administers the FCHLPM and the members of the commission are appointed by a mixture of statutory requirements and selection by the governor and state CFO.  The SBA is, ultimately, overseen by the Governor, the Chief Financial Officer and the Attorney General- all elected positions.  

Though the consulting services hired will provide their scientific evaluation to the FCHLPM who remain the decision makers, the consulting services, it seems, could be answering to the SBA.  

This adds a layer of non-transparency to the expert advising system.  In addition, it is unclear how much freedom of evaluation and access to information will be maintained by the FCHLPM.  Will the FCHLPM still be able to evaluate as they see fit or will their responsibilities be limited to decision making based only on the consulting services information? 

It is certainly not uncommon for a government agency to contract private expertise, demonstrated by by the colloquially termed Beltway Bandits.  But it is also not uncommon is for governmentally hired private expertise to produce information selectively favorable for their client.  

Clarification of who is the client helps improve transparency of potential incentives- particularly in situations of modeling science which is easily pieced together in different ways and really more theory than accepted knowledge, anyway.   

Improved transparency about how this new expert advice system will work is important for democracy and the continued integrity of the FCHLPM.  

Friday, September 19, 2014

How Different Insurance Actors Handle Recent Changes in Market Judgment of Risk


A dominant characteristic of the insurance market right now is that it does not have enough large loss to maintain pricing stability.  On Monday, the Insurance Journal ran three different stories that demonstrate several types of institutional behavioral responses to changes in the market judgement of risk.

The first article covered the recent Reinsurance Rendezvous, an annual event in Monte Carlo where the world's reinsurers get together and do whatever it is that they do behind closed doors such as, throw lavish parties.  The article reported that reinsurers are very much disgruntled with market pressure to bring their pricing down.   According to the article, originally from Reuters, Nikolaus von Bomhard, chief executive of Munich Re  claimed, “I am disappointed, exasperated, and even rather appalled by what is happening in the market."

Reinsurance shareholders are expected to benefit from the market dynamic.  The article reports, “Returning capital to shareholders reduces the pressure to do something that has higher risk,” said Moody’s analyst Stan Rouyer.  Earlier this year year, Swiss Re was reported to have done this in the form of a special dividend.

In the second article, the Insurance Journal reported that the Louisiana Citizens Property Insurance Corporation seeks to lower rates for a handful of residential policyholders and, overall, for some commercial policyholders.  The reason for the change is reported as the cost of reinsurance.   Here, the ability to offload more risk onto reinsurers for cheap has led to lower overall pricing for the public.

In the third article, the Insurance Journal reported that primary insurers (i.e. State Farm, Farmers and Allstate) in Texas are increasing homeowners rates.

The article reports the primary insurers are claiming that increased risk warrant raising rates,
Luis Sahagun, a spokesman for Farmers, said the insurer needed to adjust its rates to account for the “increasing costs associated with covering the risks faced by customers” in Texas. That includes tornadoes and hailstorms. 
“The costs associated with paying for losses resulting from fires and water damage have also kept growing,” he said.
Deeia Beck, the Texas Public Insurance Counsel reportedly argued that the raising rates are driven by profit margins,    
She said State Farm’s new rates were excessive and based on projections that “exaggerate future expected losses.” Further, the company is citing unreasonably high expenses to justify its new rates, she said.
Farmers, meanwhile, is trying to reap a “greatly excessive” profit in Texas, she said. That includes substantial profit generated from management contracts between Farmers’ holding company and its subsidiaries. 
To the extent primary insurers are able to pass perceived increases in risk onto reinsurers at a low cost, charging the policyholder for the cost of additional risk while the actual cost of managing the risk may not have changed because of the substantial decrease in reinsurance pricing would seem like a means of creating additional profit as Beck suggests.  

Here, the advantage of low reinsurance pricing is had by the primary insurer.

Though, the geophysical risk hasn't changed, pricing of the risk has changed quite a bit due to reinsurers market conditions.  Which again highlights that pricing catastrophic weather risk likely has more to do with market perceptions than scientific measures of the hazard.

Thursday, September 4, 2014

Rating Agency Outlook Performance


My understanding of a rating agencies outlook is that it is a generalized prediction about more specific predictions to be made sometime over ~2 years.  A rating is itself a prediction about a company's credit worthiness- will they or will they not pay someone.

Interestingly then, what you have with outlooks are predictions about future predictions.

How do outlooks perform?

(unless otherwise noted, a year's outlook was published the previous September.  So, the 2013 outlook was published in September 2012)

2013 Outlook









Fitch: Stable outlook over the next 12-24 months
Moody's: Stable outlook over the next 12-18 months

2014 Outlook

Fitch: Stable over the next 12-24 months; moved to NEgative outlook in January 2014
Moody's: Stable for next 12-18 months

2015 Outlook
Fitch: Negative over next 12-24 months
Moody's: In June- Negative outlook over the next 12-18 months; Can't find September report

It looks like outlook predictions don't perform all that well.  Consider Fitch's:

  1. In September 2012, they guessed stable over 12-24 months.  
  2. Twelve month's later they guessed stable again.  
  3. Three months later this became negative. 
  4. Nine months later (ie. this month) they guess negative for the next two years.  

This means that the original outlook was only good out to 15 months.  The second outlook did not hold true very long at all.  We will have to wait to see how the negative outlook of January and September 2014 will perform.

In their 2014 outlook, Fitch argued that would could trigger a downgrade in their outlook is, "Catastrophic Loss with Interest Spike,"
A sizeable catastrophic loss in conjunction with significant unrealised investment losses from an abrupt jump in interest rates is viewed as the greatest threat to the sector’s stable outlook at this time.  
I don't monitor interest rates but a sizeable catastrophic loss has not occurred (2013, 2014).  So something other than a 'catastrophic loss with interest spike' encouraged a downgrade.

This means that not only are the outlooks not so great but the reasoning behind them may not be so wholly transparent.