The study is titled "Long-run macroeconomic impact of increasing tax rates on high-income taxpayers in 2013."
In other words, this is not an immediate impact, but the "long run." You have to dig into the endnotes on page 22 to find a definition of long run: "For models of this type, roughly two-third to three-quarters of the long-run effect is reached within a decade."
Oh. So, even if one accepts the assumptions in this model — a big "if" — it still means that the loss of 700,000 jobs would not come in the first year, or by the end of Obama's second term, or even a decade from now. Yet Sessions says the jobs would be taken "from people who need those jobs," suggesting it would have an immediate effect.
Moreover, while 700,000 jobs sounds like a lot, it actually translates into one-half of 1 percent of total employment. Given that this is a long-term prediction, there is certainly a lot of room for error. So much is dependent on the assumptions in the model.
There is also another revealing endnote: "Using the additional revenue to reduce the deficit is not modeled."
That means the analysts did not even study the effect of Obama's stated purpose for raising taxes; the 700,000 figure assumes that the revenue raised from the tax increase would be used for increased government spending. Yet presumably any deal on fixing the fiscal cliff would result in a lower federal deficit, since all sides agree they have that goal.
Indeed, there are also long-term effects from permanently extending the tax cuts without cutting the deficit. This is what the Congressional Budget Office said in 2010, after studying the impact of full, partial or temporary extensions of tax cuts: "The permanent extensions of the tax cuts would have much larger negative effects in the long term than the temporary extensions because the amount of additional government debt would be so much larger."