The MBA Dilemma II – Ranking Analysis

The Financial Times does a BORDA COUNT on all the factors responsible for ranking a B-School. However, some of these factors are not really “factors of competition”. So from a voting perspective this article looks at the modified rankings of B-Schools, using the PLURALITY method of voting, considering the effect of removing the insignificant factors.

The current article will look at the intimidating number of (and hence the confusion arising due to) rankings of MBA institutes all over the world, and hence arrive at a decision as to how the rankings should be interpreted. This will help a prospective MBA to decide which schools to apply to out of so many. e.g. It will help you (“the prospective MBA” is now addressed as “you”) to answer the question “Schools A,B,C,D have similar rankings. However I can/should/have to apply to only two. Which ones?”

Which Rankings to look at and why?

Everyone rates MBA schools. The Financial Times, Business Week, Forbes, The Economist, KPMG, you, me … everyone. How credible these rankings are, is discussed in . If the rankings confuse us, would we just let them go? NO!! These experts have gone through a lot of effort when compiling the rankings and it would be a pity if we are not able to use them to our advantage. These have at least some truth in them. However, let us look at a few facts. The Financial Times Rankings are very popular. Forbes comes out with rankings bi-annually. There is a saying on wall street that when Business Week says SELL, you BUY. Schools swear by Financial Times Rankings. In fact when Indian School of Business was ranked as 20 in Financial Times this year (all facts/figures in this article are relevant to the year 2007-2008), they labelled it as a Red Letter Day. How credible were the rankings… well we will discuss that here.

So, this article will primarily deal with an exhaustive dissection of the Financial Times Global MBA Rankings. If you would rather read another ranking system, please do not read further.

STEP 1:

The first step here is to go to Financial Times MBA Ranking Page. Here you see a table with the following columns : Rank ‘08, Business School Name, Country, Weighted Salary ($), Salary Increase (%), Employment and Research. On top of this table you see a small section which says “Add more information to your table“. To the right is a sentence saying “Show all available fields“, click on it and the section expands to show you a total of a selection of 27 columns. Some are ticked and these are displayed on the table which you see below it, those that are not ticked are not displayed. You can remove any column you like by ticking the “X” on the top right hand corner of every column header or by unchecking the respective column box. All columns can be sorted by clicking on their respective headers.

27 columns, 100 schools. 2700 entries to analyse. You’re alive, I’m alive, let’s deal with it.

STEP 2:

The second step is to tick all columns so that you have a very large table. Do that now. Check all columns.

STEP 3:

You will notice that if you hover your mouse over in the section which contains all the columns OR on the header of the column, a small brownish popup will … well pop up, to tell you what the column is all about.

We shall take this description into consideration when deciding which columns to keep and which to reject. We shall proceed row-wise as in the section on top. Below I describe what FT tells us about its rankings.

3-yr Rank: These data are for information only and are not used in the rankings. Remove this column. Status: Excluded.

Aims Achieved: The extent to which alumni fulfilled their goals or reasons for doing an MBA. This column tells us, in essence, that how successful were the students in the MBA school in fulfilling their career goals. Yes, career goals. The same thing that you have been writing essays on and have been interviewed on. FT (KPMG actually) does not disclose how it collects this particular data. It could be done by interviewing alumni or comparing their career aspiration before and after the MBA. Whatever the parameters be, the result is important. Status: Included.

Audit year: Indicates the most recent year that KPMG audited a business school, applying specified audited procedures relating to selected data provided for the Financial Times MBA Ranking. <some BLAH regarding KPMG> The audit date denotes the year that the most recent selected survey data was subject to specified audit procedures. So, FT hires KPMG, an auditing company to do their rankings for them. Keep this column. Status : Included.

Careers: This is calculated according to changes in the level of seniority and the size of the company alumni are working in now versus before their MBA. Keep this column. I’ll tell you why, later. Status: Included.

Country: These data are for information only and are not included in the ranking. Keep this column. I’ll tell you why, later. Status: Included.

Employment: The percentage of the most recent graduating class that had found employment or accepted a job offer within three months of graduation. The figure in brackets is the percentage of the class for which the school was able to provide employment data. This is a reasonable column. e.g. the Employment figure for, say, Insead is 92(96) which means that Insead was able to provide employment data for 96% of its graduating class, out of which 92% had gained employment. Now since we are on an include/exclude spree we shall assume the worst. That the remaining 4% of the class for whom the employment data was not provided, were unemployedat the time. Hence multiplying the two figures we can get a good estimate of the actual percentage of employed graduates from Insead. 92*96/100 = 88.32%. Keep this column. We shall modify this for every school later. Status: Included.

International Board: Percentage of the board whose citizenship differs from the country in which the business school is located. For reasons why to exclude this column, see International Faculty. Status: Excluded.

International course: Weighted average of four criteria that measure international exposure during the MBA program. This is an ambiguous column. Simply because the terms “four criteria” and “international exposure” are not defined. Many MBA schools conduct “job search” campaigns where their students go to different countries (or one particular country) to search for a job or market themselves. This is done because there is not enough campus placement. Does this count as “international exposure”? Certainly not! Remove this column. Status: Excluded.

International Faculty: Percentage of board whose citizenship differs from the country in which the business school is located. This is an ambiguous column. FT does not provide any data as to how it uses this percentage in its formula to calculate the rank. e.g. a HIGH percentage of foreign faculty shows that the local faculty is lesser. Which means that the school is heavily dependent on foreign faculty. On the other hand it shows that the school cares enough to appoint foreign faculty to introduce variety and global perspective into its curriculum. A LOW foreign faculty might mean that the school either is very localised and lacks global perspective or is in need of funds for expensive foreign faculty and research. Ambiguity. Remove this column. On similar lines the International Board column can be removed. Status: Excluded.

International Mobility: This is calculated on whether alumni worked in different countries before the MBA, on graduation and also where they are employed today. FT does not take into consideration the country of the B-School but it does calculate a rank based on where the alumni are employed today. Yeah, outer space. There is no reason why we need this column to decide upon the school profile or our admissions decision. Eliminate this column. Status: Excluded.

International Students: Percentage of students whose citizenship differs from the country in which they are studying. FT does not tell how it uses this percentage in its calculations. Besides, fairly all B-Schools have a very diverse range of nationalities in their class. Status: Excluded.

Languages: …blah blah blah… these figures are included in calculations but are not represented on the table to avoid confusion. Case closed. Status: Excluded.

PhD Faculty: Percentage of faculty with a doctoral degree. On similar lines as International Faculty, or other common sense reasoning, this column can be excluded. Status: Excluded.

PhD Rank: <blah>. FT gives the school extra points if the doctoral graduates from that school took up faculty positions in the top 50 ranking universities. Things couldn’t get simpler. Status: Excluded.

Placements: Alumni who used the career service at their business school were asked to rank its effectiveness in their job search. Important column, keep it. Status: Included.

Rank ‘06: These data are for information only and are not used in the ranking. Status: Excluded.

Rank ‘07: These data are for information only and are not used in the ranking. However let’s keep this column, comparison sake. Status: Included.

Rank ‘08: But of course… Status: Included.

Recommends: Alumni were asked to name three business schools from where they would recruit MBA graduates. The ranking is calculated according to the number of votes for each school. In the article The MBA Dilemma I have discussed in extenso about this ranking. This is a very important statistic. Keep it. Status: Included.

Research: <blah>. FT awards points to the school if (1) the number of faculty publications is more and (2) if the faculty is still employed in the same school. Let’s get one thing very very clear. Faculty who research have very less time to take lectures, and those who take lectures don’t research. So the Research rank, in a flat flat flat world will not affect our impression of a school or the decision to join it. Status: Excluded.

Salary increase (%): The percentage increase in average alumni salary from before the MBA to today as a percentage of the pre-MBA salary. Though the terms average, before, salary and today, confuse the rating, yet let’s just retain it for now. Status: Included.

Salary Today: The average alumni salary three years after graduation. <blah> This figure is NOT used in the ranking. This is a *very* ambiguous column. Average salary of alumni three years after graduation. People take up jobs in different sectors, different levels of seniority, change so many jobs in three years and there is no wonder this figure is NOT include in the ranking. Status: Excluded.

Value: This is calculated using the salary earned by alumni today, course length, fees and other costs, including the opportunity cost of not working for the duration of the course. This is a complicated and dependent column. Retain it for now. Status: Included.

Weighted Salary: The average alumni salary today with adjustments for salary variations between industry sectors. <blah>. Lets retain it for the time being. Status: Included.

Women board: <statistical purpose> Status: Excluded.

Women faculty: <statistical purpose> Status: Excluded.

Women students: <statistical purpose> Status: Excluded.

So we have the following columns active at the end of STEP 2 :

Business School Name cannot be excluded. We have 12 columns, 100 schools. 1200 entries to compare. You’re alive, I’m alive, let’s deal with it.

STEP 4:

In this step we shall try to eliminate some more columns. But these will be eliminated, not without a side effect. Let’s see. At the end of the table there is a paragraph by FT : “Although the headline ranking figures show the changes in the survey year to year, the pattern of clustering among the schools is also significant. Some 185 points separate the top school from the school ranked 100 in the 2008 ranking. The top 15 schools, from the Wharton School of the University of Pennsylvania to the Tuck School of Business at Dartmouth College, form the leading group of world class business schools. Some 63 points separate Wharton from Tuck. The second group is headed by Yale School of Management which scored 50 points more than the Tanaka Business School of Imperial College London, leader of the third group. The fourth group is headed by Eller College of Management of the University of Arizona and includes schools ranked from 57 to 100. Some 40 points separate these 44 schools.”

The important part here is that there is clustering of schools. The clusters are Rank 1-15, Rank 16-34, Rank 35-55(56) and the last cluster is Rank 57-100. Clustering means that in each of the clusters, all the schools are more or less “equal”. FT computes the ranking according to points that it gives to schools for various values in the columns that we have been dealing with earlier. Lets make it into a table and see.

The last column is obtained by divining the Points column by the No. of Schools column. This gives an indication of the competition among schools. The more competitive the schools the higher is this number. Using no mathematics, just common sense, we can see that just 14 schools on the top separate over a point range of 63, highest in the table. Hence these schools are more competitive. The more competitive a school is the more will it take care of its students. The better are the chances of a job post your MBA. This leads us on to say that due to this kind of clustering the Rankings should be seen as clusters. In the upper clusters schools are more or less the same, in the lower ones the difference is huge.

Hence, eliminate the Rank ‘07 column. Status : Excluded.

Let’s take the Audit year column. Sorting this column from 2008 to 2002 we see that in preparing this table, data from some schools was used, which was as old as 2002! What does this mean? Is this a fair comparison? Absolutely not. Now what do we do? To make matters worse there is no correlation/clustering here. Consider e.g Rank 2 – Data collected 2004. Rank 3 – Data collected 2003. Rank 57,61,64 – Data collected 2008. No pattern at all. But we can’t do much about this. We can only trust FT enough that whatever data they collected over a vast time of 6 years was somehow balanced and weighted in a logical manner to be comparable to each other. Eliminate this column. Status Excluded.

Let’s take up the Value column. This is again a *very* ambiguous column. FT just wants to label this column as Return on Investment. However this is dependent on the country of B-School, the lifestyle of the individual, absence/presence of paid/un-paid/reverse-paid internships, the country of employment post MBA, and what not. It would just do us good to sort the values in this column and observe that there is no clustering or pattern here. e.g.

Rank 08 = 2, Value Rank = 62

Rank 08 = 41, Value Rank = 63

Rank 08 = 3, Value Rank = 73.

Rank 08 = 24, Value Rank = 100. Take a look at which B-School this is.

Eliminate this column. Status: Excluded.

Before we proceed any further let us introspect where we are going with this analysis. Now we have a hundred schools, 9 columns now, 900 entries. Now we shall eliminate some schools to make our analysis wieldly. For this we shall use clustering. We have seen above how schools tend to cluster and what it means. We shall make some assumptions.

• We want to apply to/compare 10 schools.
• The clustering phenomenon as reported by FT and as observed by us, holds true.

The Big Decision.

Now we have to choose 10 schools to which we want to apply to. I am assuming this (using the word ‘apply’), because that is the whole purpose of the article. If you wish to apply to all 100 schools, you perhaps need to reconsider an MBA.

For this purpose we will choose N, schools. But the clustering, as shown in the table above is non-linear. i.e. different number of points separate different number of schools with different ranks. Now what. Without going into a detailed analysis of how, I consider it sufficient to state that in such a case we shall choose 4 schools from the first cluster, 3 from the second, 2 from the third and 1 from the fourth. [For the curious 10*(4.5)/(4.5+2.5+1.45+.91) = 4 and so on. Round off and do the math.]

This will give us a very balanced list of schools which are very competitive and provide an effective distillation of the table. Again, the back-to-square-one question. Which ones to choose. To answer this, we note that now we have TWO rankings on either extreme of the table. The Rank ‘08 and the Recommends rank.

If it were only for Rank ‘08 things would be simple. But after eliminating many factors as unnecessary intthe Rank ‘08 column, we will stick to the more important Recommends rank. Eliminate the Rank ‘08 column. Status: Excluded. Sort the table by the last column i.e. the Recommends rank.

So now we treat the Recommends rank as the FINAL RANKING of schools. Now to choose any 10 schools. I did it randomly. Except for choosing ISB India, which has a Recommends rank of 85. And I did consider the fact that doing it completely randomly might tilt the balance in favour of USA Schools.

From 1-14, I chose 1,5,6,12. From 15-34, I chose 17,19,27. From 35-56, I chose 45,46 and from 57-100, I chose 85. [The reader can choose his/her own number of schools and in any preference.] Tick these numbers in the Recommended column and then see the first column in the table. Just before the B-School Name column there is a thin column, with a compare icon as a header. Clisk on it and the chosen (10) schools will be compared. The rest will be eliminated. This gives us a good mix of competitive schools, from different countries.

Now, let’s take up the Employed at three months column. As explained earlier the figure in brackets is the number of students for whom the data was reported. e.g. Rank =1 = Wharton the figure in this column is 93(98). For 98% of the students data was available (or made available) and out of these 93% found employment. Assuming the worst, that the 2% for whom data was missing did not find employment the employment percentage is 93*98/100 = 91.14%. Rounded off to 91%.

So instead of two figures in brackets we should have this resulting figure for a realistic analysis. Doing the math for our chosen 10 schools our table looks like this. [For the sake of keeping the width of the table within reach, I have taken the liberty to use understandable abbreviations where possible.]

Ok, now we are getting warmer. We will not eliminate these schools but try to cluster them further. We are left with 70 figures now. Right at the beginning of this article we saw what exactly each column means. Now we shall assign a grading system to each and every column, starting from the rightmost. Then we shall replace these figures with grades. The grading will be linear, irrespective of whether any value falls within it or no. This is simply for the sake of well, simplicity.

Recco = Alumni Recommended Rank.

If Recco = 1-10, Grade = 1. If 11-20, then 2, if 21-30 then 3… and so on.

Employd = Employed at three months.

If Employd = 95-100 Grade = 1, If 90-94, then 2, if 85-90 then 3… and so on.

Placmnt = Placement Success Rank.

<same grading system as Recco above>

Aims = Aims achieved Rank.

<same grading system as Recco above>

Career = Career Progress Rank.

<same grading system as Recco above> Note that Rotman which has Recommended Rank of 19 has a Career Progress Rank of 100 !! This is therefore an ambiguous column. Reason being that it measures the difference in level of seniority pre and post MBA. Now if Rotman already admits people at a senior level, will their acceleration to a yet higher level be much?No. Hence the ambiguity.

Sal(%) Inc. = Salary Percentage Increase.

If 125-130 then 1, 120-124 then 2, 115-120 then 3 and so on…

WtSal(US$) = Weighted Salary in USD.

How much difference should we take? 10K USD will be too less. 50K too high. 25K just right.

If 175K-150K then 1, if 150K-125K then 2, if 125K to 100K then 3 and so on…

Note that for each column in the resulting table you can consider the values to be Ranks of the school in their respective columns. Also since we have done a linear rating, i.e. a Rank 1 is better than say, a Rank 7, the higher the Rank the lower the school will rate. We can add a FINAL RANK column to our table adding up all the values and taking a relative subtraction from the school that ranks at number 1.

Now it is (finally!) time to look at a *very* revised version of our table. And here it is.

And looking at the results we see that even though Harvard Business School and Melbourne Business School are poles apart according to FT, by a real analysis they are 31 ranks apart. This means that our randomly chosen B-Schools did offer a healthy mix of diverse institutions. You can choose, even now to do a *very* strict analysis and render the Weighted Salary, Salary Percent Increase, Careers… (and any other column you see) as ambiguous. From here, really, it is a matter of choice to which B-School to apply to. Out of 100 B-Schools and 2700 parameters we have distilled 10 schools with one FINAL RANK parameter. This should end most of the confusion surrounding any kind of ranking system. This article might not be *very* representative of B-School profiles, but gives a much clearer picture on

• What parameters should a B-School be evaluated on?
• What is the actual situation in Rankings?
• Are any ranks/representations ambiguous?

I hope that to the curious reader I have answered these and many more confusing questions. Feel free to comment! Somebody… I need a shot or two…