A pandemic of the size you asked about is a credible threat. I've seen articles over the last couple of years where researchers attempt to model the spread of diseases and what affects how quickly they spread. The US CDC spends a lot of time thinking about this kind of thing. Given the diversity of research institutions who are working on understanding the spread of disease, I think the threat of a pandemic is a very credible.
[Update for Oct 2021]: And COVID-19 proves my point about pandemics as credible threats. If COVID were a bit more lethal, we'd be in real trouble with much higher casualties. Also, my original post didn't account for the effects of politicisation of the pandemic and those who will prioritize something other than making the pandemic end as soon as possible.
The problem with short gestation period diseases is if you, as the disease, kill your host too quickly then you don't get a chance to spread yourself around as much. Diseases that kill quickly typically turn into diseases that don't kill too quickly because a fast-killer doesn't "survive" long in the population while tamer mutations have more time to expose other hosts. A virus will change over time in relation to the selection pressure of a host's immune system (herd immunity may play a factor too). The longer a virus survives and the more hosts it infects, the greater the chance for mutation.
There's this article from 1999 that talks about mutation rates in RNA viruses. (I'm sure there are newer, better articles to be had.) The mutation rates are stupid high on the order of around 75% per genome per replication. The mutation rate is much lower for DNA-based microbes (including both viral and cellular organisms) at .34%. The authors of the paper state that these numbers are hard to gauge because the mediocre quality of the data. If that's just one nucleotide involved in the mutation, at trillions or quadrillions of replications across many different hosts, there will be mutation.
Older Answer to the Question
The ideal pandemic agent might have the following characteristics: Make the virus airborne, highly contagious (requires very small doses for infection), contagious shortly after infection, and a long gestation period in the months or years range before it induces acute illness/death. Basically, airborne HIV.
Virus based diseases such as HIV, Ebola or influenza have historically been very difficult to study based on how long it takes for effective treatments to appear after a virus discovery. We've known about HIV since 1980 but it took years to develop effective treatments. If this is a brand new disease then it will take a while for medicine to learn how it works and develop an effective treatment thus giving the disease plenty of time to spread. Ebola treatments are relatively new but medicine has known about it for decades (the Hot Zone was published in 1994). While the pandemic doesn't strictly need to be virus-based, having it be so decreases the changes of rapid effective analysis/treatment by the CDC or local equivalent.
Let's assume the following characteristics of this virus:
- Average gestation period: 6 months, standard deviation of 2 months.
- Illness Rate: 50%
- Fatality Rate: 10% (per your question)
- Required dosage for infection: 100 viruses.
- Time to contagious: 3 days, standard deviation of 1 day.
- Symptoms: Flu-like, coughing, aching, runny nose, headache, fever.
Plausible Scenario:
Disease originates in sub-Saharan Africa (sorry Africa). Patient Zero catches the disease from infected pig meat. The virus causes minor inflammation in the nasal cavities, thereby inducing sneezing. It is passed to a foreign aid worker who is about to finish their tour and go home to California after stopping off in London and Paris to visit relatives. The aid worker sneezes on the plane without covering her nose (shameful). The water particles from her sneezes spread the virus to everyone on the flight. In 2 to 4 days, all those travelers will infect the people they know. Even if a person sneezing in a huge airport like Heathrow only infects 10% of the people moving through then that's thousands or tens of thousands of individuals infected. And, people flying through Heathrow go literally everywhere on the planet. People working in the airport will contribute greater still to the infection rates as they have a persistent presence to keep recontaminating the airport.
After stopping off in Paris and London, she heads home to California. The virus is still undetected and unknown. There are no quarantines for infected individuals, no blockades of infected countries. Assuming she has layovers in JFK in NYC and arrives in LAX, two of the largest airports in the US are now infected.
I think you can see where this is going. If nothing else, this kind of a disease is a special threat to the First World because of how interconnected those countries are and the common use of air travel. The power of the CDC can't help you if the virus is unknown.
Compound the effects of this new virus by having the virus start spreading in May/June/July. The 6 month gestation period will coincide with flu season and that will complicate treatment because doctors will think that people are getting sick from the flu, not this new virus. It will appear to the public as a terrifying flu season because people are getting sick and dying from the flu or something like the flu but the flu shots aren't working. The usual methods of quarantine for flu patients won't work because practically everyone was infected months earlier.
What makes this scenario plausible is the high degree of interconnectedness in the First World because of air travel. It's an epidemiologist's worst nightmare to have a lethal, airborne virus that gets spread around by air travelers coughing and sneezing on planes.