As I mentioned, I have been reading Social Security, by Jagadeesh Gokhale. He has developed a forecasting model for Social Security that is more fine-grained than the one used by the system’s own actuaries. Some excerpts below.p. 29:

Future growth of aggregate “effective labor inputs” will be diminished significantly because of declining labor quality…The projected decline in effective labor inputs results from the interactions of demographic and economic forces over time.

Like Goldin and Katz, Gokhale sees a slowdown in the growth of educational attainment. However, Gokhale sees this as coming not from a policy failure but from the nature of demographic change.
p. 51:

although education rates for all population groups may increase over time, changes in the population weights of different population groups will also determine the population’s average educational attainment in future years.

In a sense, he is suggesting that the U.S. will suffer from a Gregory Clark effect in reverse. Clark argued that prior to the onset of the industrial revolution, British upper classes reproduced at a higher rate than lower classes. Gokhale uses careful wording, but what I see him suggesting is that Social Security’s outlook will be affected by the higher rates of reproduction of the relatively less-educated portion of our population.

The Social Security Administration does not undertake such fine-tuned demographic analysis. As a result, their forecast for Social Security’s finances, while sobering, is not as pessimistic as Gokhale’s. On p. 124, Gokhale presents a table with the comparison. For a 75-year time horizon, SSA projects a Social Security deficit of 1.7 percent of payrolls, meaning that it would take a 1.7 percentage point increase in payroll taxes, starting now (actually, starting three years ago) and continuing indefinitely, to close the deficit. Gokhale, in his baseline forecast, sees a deficit of 3.35 percent of payrolls, meaning that it is twice as large.

p. 144:

It now appears more appropriate to treat and evaluate Social Security as a public retirement savings program rather than as a purely social insurance program.

The point here is that when the age of eligibility for benefits was not far from average longevity, Social Security acted like a program to insure people against the risk that they would outlive their savings. Today, however, with an age of eligibility well below the age of longevity, Social Security acts more like a government-run retirement savings system.

There is much more in the book, including evaluations of different Social Security reform proposals. From now on, I would suggest that any pundit who makes pronouncements about Social Security deserves no credence unless he or she has read Gokhale’s book.