Saturday, February 28, 2009

Good schools and bad schools

People often talk about good schools and bad schools. But how are these defined. In practice a school is considered good if its students do well on standardized achievement tests, a school is considered bad if they don't. So what determines how students do on standardized tests. Empirically in the United States systematic differences in average student achievement between schools appear to be almost entirely due to differences between the students attending each school as opposed to differences between the schools themselves.

For example suppose school A is currently considered good because its students do well and school B is currently considered bad because its students do poorly. And suppose we were to start busing all of A's current students to B and all of B's current students to A. What would we expect to happen? We would expect the students who had been doing well at A to continue to do well while attending B and the students who had been doing poorly at B to continue to do poorly at A. In other words A would now be considered a bad school and B would now be considered a good school although the teachers, administrators and physical facilities at A and B would not have changed.

Unfortunately most public debate regarding educational policy totally ignores the above giving it a surreal quality. Democrats advocate improving schools by spending more money on teachers and facilities, Republicans advocate improving schools by breaking the teacher's unions and instituting vouchers. However there is no reason to believe either plan will make any significant difference in student achievement.

The Coleman report found that schools make little difference. Kansas City is an example of the failure of vastly increased spending to improve schools.

Friday, February 27, 2009

Early retirement and social security

Suppose you retire early. How much does this reduce your eventual social security benefit? Generally not as much as you might expect if you were well paid. This is because social security is front loaded, in other words you tend to earn a disproportionate share of your benefit in your first years of work. Traditional private pension plans are often back loaded.

Social security benefits are based on your average monthly earnings (from jobs that paid into social security) over your 35 highest earning years prior to age 62. When computing your average monthly earnings, your earnings are scaled to account for inflation. So for example if you were born in 1947 (so you would be 62 in 2009) your earnings in 2008 are multiplied by 1.00 but earnings from 2000 are multiplied by 1.26, earnings from 1990 are multiplied by 1.92, earnings from 1980 are multiplied by 3.23 etc.

Your average monthly earnings are then converted to a monthly social security benefit via a formula that gives greater weight to the first dollars earned (thereby favoring low income workers). So for example if you were born in 1947 your benefit will be 90% of your average monthly earnings up to $744 plus 32% of any additional average monthly earnings up $4483 plus 15% of any average monthly earnings exceeding $4483. The formula for other dates of birth will have different break points but the same general nature, your first dollars earned will count for 6 times as much in figuring your benefit as the last dollars earned (assuming you earn enough to get your average earnings above the second break point).

So in my case I can make the following rough estimates. I currently have 26 years of maximum earnings. If I hadn't been laid off I could have expected another 6 prior to age 62 (I started working too late to get in the 35 years of maximum earnings required to get the maximum benefit). So my current average monthly earnings are about 26*102000/(35*12) = 6314 (note 102000 is the 2008 earnings limit) which corresponds to a monthly benefit of .90*744+.32*3739+.15*1831 = 2140.73. Another 6 years of maximum earnings would increase my average monthly earnings by 6*102000/(35*12) = 1457 and my monthly benefit by .15*1457 = 218.55. Hence increasing my average monthly earnings by 23% by working another 6 years would only increase my benefit by about 10%. This is a very rough estimate (for example each year of maximum earnings does not actually contribute exactly the same amount to your average monthly earnings) but illustrates the point.

Note the combined social security tax rate is 12.4% so each additional year of maximum earnings represents $12648 of additional tax paid into the system. Over 6 years this is about $75000 for an added monthly benefit of $218 not a terrific deal.

See here for more on how social security is calculated.

Added 1/12/2014:  The general point of this post, that social security is front loaded,  is correct but a detail is wrong.   Your 35 highest earning years which are used for computing your benefit can include years in which you are 62 (or older).  So by continuing to work I can increase my benefit by a bit more than indicated above.  See here for more.

Thursday, February 26, 2009

Last day, First post

Today was my last day at work. I was laid off after 25 years with a large American technology company. Fortunately I will not be under any immediate financial pressure as (barring a total economic collapse) I will have enough retirement income to live on if I am careful. So I have a lot of options.

One of which is to start a blog which I am doing with this post. How hard can it be? I will post about things that interest me, politics and public policy, CO2 and climate change, issues facing recently laid off fifty something single white guys etc.