ANALYZING ACADEMIC STAFF PROMOTION CRITERIA IN NIGERIAN

UNIVERSITIES USING A NONERGODIC MARKOV CHAIN MODEL

P. O. Ekoko
Abstract
In this paper promotion criteria for some Nigeria Universities are presented.
Based on certain assumptions, a Markov chain model that is not
irreducible has been formulated. The model is applied to analyze two
aspects of promotion criteria in Nigerian Universities namely: the
expected number of years an academic staff can stay in a cadre in
which he was initially employed and prediction for future years the
percentage of academic staff in each cadre. These analyses were
numerically carried out using the data obtained from one of the
universities.

Keywords: Markov chain, promotion criteria, irreducible Markov chain, ergodic.

Introduction
In Nigeria, a university is established and owned by either, the Federal Government, State
Government or Private individuals including religious bodies. Though the National Universities
Commission (NUC) is the body that regulates the academic standard and accredits degree programmes
in all Nigerian Universities, the commission does not formulate promotion criteria for academic staff in
any University. Academic staff promotion criteria are set up and approved by the Governing Council of
the respective Universities. And this is the reason why academic staff promotion criteria differ from one
University to another. For example, as can be seen in UNIBEN ASUU (2004), that is Fig. 1, ten papers
are among the requirements for promotion of academic staff to the rank of professor in two Federal
Universities namely: University of llorin, Iiorin and Ahmadu Bello University Zaria while twenty seven
papers are among the requirements for promotion to the same professorial rank in the University of
Benin City, which is also a Federal University. More details about the twenty seven papers required by
the University of Benin can be found in UNIBEN
(2003) .
It could be remarked that there is only one University System in Nigeria that produces
graduates for the same labour market, yet the number of publications for promotion of academic staff at
the various levels in some Universities are undoubtedly lower than the number in some other universities
be it Federal, State or Private. In fact, Moye (2004). puts it this way. “To the average Nigerian, a
professor is recognized as such irrespective of the standard of attaining the position or the University that
gave the chair. This kind of reasoning has constantly brought accusations and counter accusations
between Universities Management and their corresponding branches of labour unions, Academic Staff
Union of Universities (ASUU). While University Managements may argue that the work force is
becoming top heavy with too many professors evolving at a rate much higher than financial increment
from government subventions, ASUU believes that the present duties and number of students an
academic staff has to cope with has been on the increase in geometric proportion for over ten years now.
The unions wonder why those who became professors with less than half of the present number of
publications should confiscate or throw away the ladder with which they climbed to the top so that others
do not find the means of reaching them. There is hardly any ASUU branch union and its University
Management that has not had conflict due to this “publish or perish syndrome.” Many of such cases have
resulted in union strikes before settling issues through dialogue.
 Comparative Analysis of Promotion Criteria of Some Nigerian Universities
123 45 67 8 9 10 11
Post UNI .ABU UI UNI UNI UNN (1 NNA. UNAAB UNIABUJA OAU UNIBEN
LAG PORT LORIN paper = 0- AZIK IFE
3 Pts) UNIV
Asst. NA 3 Conf. 1-2 Papers NA NA 15 Points NA Ph.D, with No. Paper, 1 Journal NA Ph.D or M. Sc 1
Lect. Papers/Sem MSc. With 1 Journal Paper (5 Journal Paper
To Paper or 2 Edited points)
Lest. Conference Papers’
II
Lect. NA 3 Jour. 5-7 3 Jour. 4 Jour. 25 Points 4 pts Ph.D with 4 Journal Ph.D with 2 4 Papers Ph.D. or M.Sc. 3
II To Papers for Papers Papers (12 Papers or 3 Journal Papers (8 Publications 2 must
Lect. I Non-Ph.D 2 Ph.D is Pts) Papers and Edited Points) be JoumaLs
Jour. Papers/ Desired Conference Papers M. M.Sc. with 3
or 3 Conf S. with 3 Journal Papers or (12
Papers for Papers and Edited Points)
Ph.D Conference Paper.
Lect. 15 Ph.D 10-15 Papers 17 Jour. 7 Jour. 30 Points 10 Pts Ph.D with 10 Papers. Ph.D with 5 7 Papers Ph.D or recognized
1 To Points Desired 6 Depending on the Papers or Papers (20 6 of which must be papers or (25 Professional
Sen. Jour. Paper 5 faculty e.g. Social 20 Points Pts) Journal Papers or 11 points) M. Qualification. 8
Lect Conf Papers. Science which also Papers. 5 of which Sc. With 7 Publications. 6 must
emphasizes will be Journal and 6 papers (35 be Journal Articles
Foreign* Conference Papers points) plus 12 pts
Component from S.
Sen. 21 Ph.D 10 L.
15-20 Papers, 25 17 Papers 8 Jour. 60 Points 20 Pts Ph.D with 5 Journal Ph.D with 9 8 Papers Ph.D or Recognized
Lect Points Jour. Papers Papers (e.g. Social (50 Papers (24 Articles with i I in papers (45 Profession «
To 5 Conf. Science) Points) Pts) Print and 4 in Press or Points) Qualification. 18
Assoc. Papers (15 Depending on 14 Journal Articles, 2 Publications 75% in
Lect. Public) Faculty (General, Conference Papers or Print. 3 in 2
best 10 that reflect 13 Journal Papers and Different Indexed
the totality of their 4 Conference Papers Journals 3 Refereed
Contribution to (75% in print) Proceedings Plus 24
Scholarship in their pts.
Key : NA means information notDiscipline)
available
 1 2 3 4 5 6 7 8 9 1 0 1 1
Post UNI ABU UI UNI UNI UNN (1 NNA. UNAAB UNI ABUJA OAU UNIBEN
LAG PORT LORJN paper = 0- AZIK IFF
3 Pts) UN IV
Assoc. 24 Points Ph.D 22 22-25 Papers (e.g. 22 Papers 10 Ph.D 30 Pts 20 Journal Papers, 15 Ph.D with 12 10 Ph.D 27
Prof. To Public i.e. Agric, Art): 30-40 or 65 Journal Requi. must be in Print or 19 Papers or 60 Papers Publications.
Prof. 15 Jour Papers (e.g. Social Points Papers (65 pts) at Journal Papers and 2 Points 75% in Print. 5 in
Paper & 7 Sciences with (28 pts) least 15 Conference Papers or 3 Indexed
Conf Paper emphasis on Pts via 18 Journal Papers and Journals. Not
Foreign) Jour. 4 Conference Papers more than 3
Component of which 75% in Print Proceedings Plus
Depending on 30 Pts
Faculty. (Generally,
best 10 that reflect
the totality of their
Contribution to
Scholarship in their
Ph.D Prefer. Desired discipline)
NA Desired NA Required NA Required from S. L. Required feom 1 or 2 Required from S.
Higher form S. L. from L. 1 from A. P. but Six Years of Grace A. P. extra L.
Deg. e.g. Period from 2003 Papers
Ph.D or without
its Ph.D
Profess from A.
Equiv, P.
after
Good
Hon.
Deg.
Foreign' Nil Nil *(Please see below Nil Nil Nil Nil 15% and 20% for A. P. Nil Nil Required plus
Index and Prof. Spread
Publ/ j Respectively; Spread
Spread Not Required !
* Ibadan: Each faculty should determine which Journals are learned Journals and which are not. What is more important is the contribution of the candidate to
scholarship than the Journal in which it is published. It is not to make a distinction between local and international Journals: for a Journal, which is local in one
place, may be international in another.
Fig 1: Comparative Analysis of Promotion Criteria of Some
Nigerian Universities. Source: UNIBEN ASUU (2004).
 Strikes are counter productive and often evolve from a critical, strained labour relation. Many
research work including Lilien and Rao (1975), Carey and Sherr (1974), and Grinold (1976), on
Manpower Management are focused on labour relations especially the aspect of resolving conflicts that
arise between management and union. The scientific research papers often use models of Markov chain as
could be found in Schachtman and Hogue (1976), Simmons (1971), and Villiant and Milkovich (1977).
Markov Chain models are stochastic processes where the processes move from one outcome (or
realization) to another outcome according to set of probabilities. Stochastic processes are collection of
random variables whose values are observations at certain points in time. Such processes are very general
in nature and some specific assumptions must be imposed in order to yield a tractable model. Specific
illustrations of such tractable models can be found in Dinkel et al (1978), Hillier and Lieberman (2001),
and Taha (2002). A Markov Chain has been derived in the following section to assist in analyzing the
problem of academic staff promotion criteria in Nigerian Universities.

Nonergodic Markov Chain Model

Before discussing the model, it is necessary to define some terms as follows:
Definitions
a. Irreducible: A Markov is said to be irreducible if it is possible to start from state i and reach state
j in some finite number of transitions \/ states i and j i.e. Pij(n) > 0 for some n, V1 and
j-
b. Periodic: A state, i is said to periodic with period t> 1 if it is possible for the chain to be in state i
only for multiples of t. in other words pi/n) = 0, whenever n is not a multiple of t. If t = 1, the
Markov chain is said to aperiodic. The idea of periodicity relates to a cyclic nature with which
the process returns to certain states.
c. Ergodic: If a finite markov Chain is irreducible and aperiodic then it is ergodic. Ergodic Chains
are important because for such a chain it can be shown that the long-run probability exist and are
uniquely determined. That is
Lt Pij(n) =πj, j = 1(1 )A
n —>oo
where the πj satisfy the Steady State Equations

These probabilities, nj are called steady state probabilities for they are the probability of finding the
process in state j after the process has gone through a large number of transitions.
d. Absorbing State: A state is said to be absorbing if whenever you enter it you cannot come out of
it. That is pii = 1. If at least one of the states in transition matrix is absorbing then the Markov is not
irreducible and it cannot be ergodic. It is not possible to complete the long-run probabilities icj (j = l(l)N)
for a nonergodic Markov chain.

Assumption of the Model

The following are the assumptions of the model:
a. The transition matrix ensures that an academic staff who is due for promotion
can only move to the next rank at the one step transition.
b. Demotion is not allowed.
c. Each time period is one calendar year.
d. The stationarity property of Markov chain holds.
 e. Let the transition matrix of order /V (having two absorbing states) of the model that satisfy the
above assumptions be denoted by:

and Pj" = Pj"> is the probability of an academic staff remaining in cadre / after N years. It can be
proved that for the transitions matrix (5) of the Markov model in which demotion is not allowed the
expected number of years an academic staff can remain in cadre j given that he started in cadre j is
denoted by

f. Apart from computing the expected number of years an academic staff can
remain in a cadre using equation (7), the noncrgodic Markov model can also be used to aid in predicting
(he number of academic staff perso each of the cadre in future years. This is explained as follow: If
at time t,
g. there are nti academic staff in cadre i ,i < j with Py probability of moving from cadre i to cadre j,
then the total number of academic staff in cadre i at time (/ + 1) is denoted by

and Hj in (a) is the number of academic staff that are constantly recruited into cadre / annually.
If recruitment is done only at eh lowest cadre / = 1, then (a) of equation (8) is used for only the first
cadre while (b) is used for other cadres /< 2. However, if each cadre j has a constant number ttj of staff
that is recruited into the cadre then in place of equation (8) we use equation (9) as follow:

h. Using the same transition probability matrix, we only need to substitute n (t+)j for n(t)l to
obtain nj(l +2) and so on.
Using the same transition probability matrix, we only need to substitute n (t+)j for n(t)l to obtain nj(l +2)
and so on.
.Model Application
Using five-year data obtained XYZ University in Nigeria, the transition matrix in (10) was obta
ved. The actual name of the University in Nigeria is being withheld because of the confidential
consi
derat
ion
of
the
Univ
ersit
y.

(10)

(10) is the partitioned transition matrix of Academic Staff Promotion in XYZ University. In collecting the
data from XYZ University, the various ranks of academic staff were classified and denoted as follow:
Ju - Junior lecturers made up of Asst. Lectures, Lecturers 11 and Lecturers 1.
Sn - Senior Lectures As -
Associate Professors
LBBRD - Those who could leave by resignation/dismissal LBRC -
Those who could leave by retirement/contract appointment.
Using equation (7), the expected number of years to remain at each of the first three cadres, for
each starting cadre j namely; Junior Lectures (j - 1), Senior Lecturers (j = 2) and Associate Professors (J =3)
are 3.08 years, 3.08 years and 22.44 years respectively. By this model someone who enters the academic
profession newly as Associate professor is expected to remain at that cadre without promotion for 22 years.
It implies that the model discourages someone entering the profession newly from top without commensi
raT experience which is very vital in the teaching profession. This shows that it is better to start the
academic profession in the XYZ University from the lower cadre and acquire the experiences gradually.
The model’s use in predicting the number of academic staff in each cadre in future years is
numerically illustrated for XYZ University as follow: XYZ University presently in 2006 (considered as the
zero year) has 930 (62%) Junior Lecturers, 270 (18%) Senior Lecturers, 180 (12%) Associate Professors,
and 120 (8%) Professors. With up to 48% of its academic staff being Senior Lecturers and above, the XYZ
University was able to have all its degree programmes accredited by the National University Commission
(NUC) last year. Consequently, the present policy of XYZ University is to employ academic staff only into
the Junior Lecturer cadre. And the policy fits into the equation (8) of the Markov model.
Starting with the above present number of academic staff in each cadre with constant
employment of 30 academic staff annually into the Junior Lecturers cadre, the number of academic staff in
each cadre at the end of the first year is given using equation (8) as follows: 560 (44%) Junior Lecturers,
405 (31%) Senior Lecturers, 192 (15%) Associate Professors and 124 (10%) Professors. At the end of the
first year the total number of academic staff observed in XYZ University is 1,281 instead of 1530. the
difference of 242 is the number of academic staff that left the XYZ University through the two absorbing
states.
Continuing this way, the results for the first five years are recorded in Table 1.
Table 1: Showing the percentage of academic staff to be in each cadre of XYZ University in the first. live
years with constant annual employment of .30 academic staff into the Junior Lecture cadre.
Year Ju (%) Sn (%) As (%) Pr (%) Expected Observed No. lost
No. in the No. in the through
Year year Absorbing
states
2006 62 18 12 8 1500 1500 0
2007 44 31 15 10 1530 1281 249
r h
2008 32~ 35 21 12 1311 1084 227~
2009 25 34 27 14 1 114 922 192
2010 20 31 31 18 952 788 164
2011 18 26 35 21 818 690 128

Recommendations
The second assumption of this model states that demotion is not allowed, which was exactly
what happened for the years in XYZ University. However, in many real situations both promotion and
demotion can occur in a given year. That is in annual appraisal exercise some staff may experience
promotion or remain in their level if they have not met the requirements for promotion while others may
be reprimanded by demotion. By modifying the second assumption of this model to include cases of
demotion and possibly double promotion the model can be made more relevant and applicable to many
establishments. These cases can be good extension of this research work.
Conclusion
It should be encouraged that the nonergodic Markov chain model be applied to determine the
expected number of years an academic staff employed at a certain cadre will remain in that cadre. This
was illustrated using XYZ University. The second use of this model is in determining the percentage of
academic staff in each cadre in future years. In (b) of equation (8) the quantity y p to is made up of those
who after the promotion exercise were not promoted to cadre / and those / ''' that were promoted from
other lower cadres i, (i < j) to cadre j. As can be seen in Table 1, in the long run (despite the Hj constant
addition to the Ju cadre) the higher cadres tend to be top heavy because more of the academic staff
would have been promoted to these cadres which is the reverse of the initial spread. Also in the long run
the number of academic staff that will resign, be dismissed, retire, or take up contract appointment (i.e.
to be found in the absorbing states) will reduced.
References
Carey, K. J. and Sherr, L. A. (1974). Market and Price Factors in Transaction -to Transaction Price Change
Behaviour of Common Stocks, Applied Economics, Vol. 6.
Dinkel, J. J.; Konchenberger, G. A and Plane, D. R. (1978). Management Science Text and Applications.
Homewood: Richard D. Irwin, Inc.
Grinold, R.. C. (1976). Manpower Planning with Uncertain Requirements, Operations Research, Vol. 24.
Hillier, F. S. and Lieberman, G.J. (2001). An Introduction to Operations Research, 4,h Edition. San
Franscico: Holden Day.
Lilien, G. L and Rao, A. G. (1975). A Model for Manpower Management. Management Science, Vol. 21,
No 12.

Moye, A. S. (2004). Further Rationale for ASUU-UNIBEN Proposal on Promotion Criteria. A Write
up from UN1BEN ASUU to the Chairman, UNIBEN Governing Council and Signed-by A. S.
Moye, Chairman, UNIBEN ASUU.
Schachtman, R. H. and Hogue, C. J. (1976). Markov Chain Model for Events following Induced Abortion,
P. 0. Ekoko
Operations Research, Vol. 24.
Simmons, D. M. (1971). Common-Stock Transaction Sequences and the Random-Walk Model, Operations
Research, Vol. 19.
Taha, H. A. (2002). Operations Research: An Introduction. 2'ui Edition. New Jersey: Prentice Hall Inc.

UNIBEN ASUU (2004). Comparative Analysis of Promotion Criteria of Some Nigerian Universities. A
Document of UNIBEN ASUU sent to Members of Senate of the University of Benin.
UNIBEN (2003). Regulations Governing the Service of Senior Staff, University of Benin. A Handbook
Publication of lie University of Benin on Regulations Governing the Service of Senior Staff of the
University of Benin, Benin City.
Valliant, R. and Milkovich, G. T (1977). Comparison of Semi-Markov and Maikov Models in a Personnel
Forecasting Application, Decision Sciences, Vol. 8.