Grade inflation - a mark of human kindness
Why has the grade point average of American college students, like English A-level scores, been rising? Why has executive remuneration outstripped the earnings of other employees? Why have audit standards been falling? The answer to all these questions is the same. Grade inflation occurs whenever one group of human beings is asked to comment on the performance of another.
The Lake Wobegon effect is named after Garrison Keillor's mythical prairie town where all the children are above average. Most of us are good-natured and want to think well of our fellows. And it is good that it is so. Dedicated instructors sympathise with their students, effective boards support their chief executives, good accountants not only count the profits but make an effort to understand the business. I have never heard of a remuneration committee that thought its executives should be paid in the bottom quartile of the distribution. And I hope I never will: if that is their opinion they should pluck up courage to fire them. But no complex maths is required to see that if people are picking numbers that are, on average, above the average of everyone's pick, that average will steadily increase.
The second component of grade inflation is Goodhart's Law. When monetary targets became fashionable, economist Charles Goodhart observed that when an aggregate became a target its significance immediately changed. Any specific indicator adopted as a measure of performance will improve relative to overall performance. If hospital managers are encouraged to ensure that no one has to wait for more than six weeks, few people will wait for more than six weeks but a lot may wait for 41 days. If executive pay is related to EPS targets, earnings per share will tend to grow even if the long-term value of the business shrinks. Accountants will fit their reporting to generally accepted accounting principles. Students will focus on the particular skills on which they are measured.
So when people ask whether the performance of students has improved or the standards applied to them have declined, they pose a good question but a difficult one to pin down. Isaac Newton was a great physicist but he would have flunked A-level. There is so much he did not know. It is a tougher job to discover the laws of motion than to remember them, but it is perfectly sensible that today's students are given the latter task rather than the former. When grandparents complain that young people cannot do long division, and grandchildren that old people do not know how to reboot the computer, both groups are right. People learn to do the things that are required of them, and so it is with examination performance. The only antidote to Goodhart's Law is the balanced scorecard - use different indicators and change them frequently.
Grade inflation is also the product of competition. Competition improves performance and mostly this is good: it leads to lower prices and shorter queues at the checkout. But the process has perverse results when the product is performance measurement, and the buyer is the person whose performance is being measured.
If Professor Nice gives all his students As, and Professor Nasty gives all her students Cs, then students will prefer to enrol with Professor Nice. Not only does this increase the average grade, but it puts pressure on Professor Nasty to conform. Few executive remuneration consultants win business by being tough on the people who hire them. Only in the aftermath of Enron was objectivity a selling point for auditors. What customers want from examiners and advisers is the appearance of rigour but not the substance.
Grade inflation is the inevitable product of benign human nature, of pleasant social interactions and of self-interested commercial ones. But some institutional structures invite grade inflation and others resist it. Universities are lax examiners of their own students but tough examiners of school students. Privatised examination boards produced an improvement in scores, just as competition among auditors led to higher reported earnings. And the creation of remuneration committees produced an upward spiral in executive pay.
分数膨胀”是人性弱点
为
何美国大学生的平均分数，像英国的A级考试一样，一直在提高？为何高管薪酬的增幅超过其他员工？为何审计标准一直在下降？所有这些问题的答案都是一个。只要要求人类的一个群体去评价另一个群体的表现，分数膨胀就会出现。
沃比根湖效应(Lake Wobegon)的名称，源自盖瑞森?凯勒(Garrison Keillor)虚构的草原小镇，那里所有儿童的智慧都高于平均水平。我们多是好心人，愿意往好处想我们的同类。这很好。敬业的教员体谅学生，有效的董事会支持他们的首席执行官，优秀的会计不仅记录利润，而且还努力了解业务。我从未听说有哪家公司的薪酬委员会认为，其高管的薪酬应定在低端水平。我也希望自己永远不会听到有这样的薪酬委员会：如果他们真有这个观点，就应该鼓足勇气裁掉这些高管。但不难推理，如果人们普遍按高于平均的标准挑选数字，那么这个平均标准就会逐渐提高。
分数膨胀的第二个成份，是古德哈特定律(Goodhart’s Law)。当货币政策目标开始风行时，经济学家查尔斯?古德哈特(Charles Goodhart)观察到，当总数成为目标时，其重要性会立刻发生变化。任何作为衡量标准的具体指标，相对于整体水平都会得到提高。如果有人鼓励医院管理人员保证等待时间不超过6周，确实会很少有人等待超过6周时间，但很多人可能要等上41天。如果高管薪酬与每股盈利(EPS)目标挂钩，即使企业的长期价值缩水，每股盈利也会有所提高。会计师会把他们的报告套上公认会计准则。学生将关注自己需要接受考核的技能。
因此，如果有人问：是学生的成绩提高了，还是应用于他们的考核标准下降了，这确实是个好问题，但答案难以确定。艾萨克?牛顿(Isaac Newton)是一名伟大的物理学家，但他的A级考试很可能不及格。他不具备很多知识。发现运动定律比记住它们更难，现在的学生要做的是记住而不是发现定律，这完全合理。祖父母抱怨年轻人不会做多位数除法，孙辈抱怨老人不会重启电脑，两代人都没有错。人们只会学习要求他们学习的东西，考试成绩也是如此。弥补古德哈特定律缺陷的唯一方法，就是平衡记分卡(balanced scorecard)，采用多种指标，并经常更换这些指标。
分数膨胀还是竞争的产物。竞争会提高表现水平，这在多数情况下是好的：有助于降低价格，缩短收款处的长队。但当产品是业绩考核，而考核对象是购买者的业绩时，这一过程就会造成异常的结果。
如果“好教授”给所有学生的分数为A，而“坏教授”给所有学生的分数为C，学生就会偏好选“好教授”的课。这不仅会抬高平均分数，还会给“坏教授”施加压力，迫使他向平均标准看齐。高管薪酬咨询顾问如果苛刻对待聘请他的人，自然不利于自己的业务。只有在安然(Enron)事件后，客观性才成为审计师事务所的卖点。客户希望从监察者和顾问那里得到的，是表面上的严格，而非实质上的严谨。
分数膨胀是人类仁慈天性的必然产物，也是令人愉快的社交和利己主义的商务往来的产物。但有些机构的架构滋长分数膨胀，有些则抵制这种倾向。大学对大学生的考核标准宽松，但对报考的中学毕业生考核严格。私有化的考试委员会将提高分数，正如审计师事务所之间的竞争导致高报赢利一样。薪酬委员会的设立，会导致高管薪酬扶摇直上。
