We have taken note of the alternative costing of the NPP Free NHS policy
produced by the Center for National Affairs (CNA).
This is a very welcome development, since it heralds a new era of
competitive policy-based political debate in this country.
In the specific case of the CNA?s intervention in the SHS debate, however,
there are simply too many gaps in their analysis to allow for its use as
an effective tool in interpreting the financial impact of the proposed
fee-free SHS policy.
We however welcome their suggestion to use primary Ministry of Education
data and the results of the Ghana Living Standards Survey ? round 5.
Indeed, we have overcome our earlier concern about the Ministry?s dataset
not being up to date by using careful projections. This has enabled us to
follow an identical approach to the CNA?s, and also provided further
corroboration of our earlier analysis. We are now proceeding to publish
We had chosen to use JHS enrolment data in our earlier analysis because
the treatment is smoother that way. The critical impact of the policy is
derived from its objective of ensuring seamless ?transmission? from JHS to
SHS. In that light it makes sense to focus on that level and to ?stretch
out? the average per capita costs to accommodate the analysis. That
notwithstanding, the approach used by the CNA, provided the shortfalls in
their methodology are remedied, should lead to outcomes similar to those
produced by the original IMANI approach.
We will now proceed to identify some of the most worrying gaps in the CNA
1. CNA claims that according to GLSS data the amount spent by households
on each secondary student is $148. This is odd since the actual average
amount indicated on page 124 of that document is GHC244. This figure is
converted to $148 by the CNA, wrongly.
The CNA should have used 2008 constant dollar rates rather than the early
2012 dollar-cedi conversion rates that they used. Had they done so, the
figure would have come to: $256. Indeed, the GSS data dates back to 2005,
so a truly robust treatment would have projected to 2011 in Ghana Cedis
before commencing the dollar conversion. Again, the final sum would have
been more than twice what the CNA claims.
2. The CNA assumes that the total cost of catering for the expanded intake
of students is limited to Government?s absorption of household
expenditure. They do not account for the fact that for each new student
beyond fixed enrolment, the existing costs ALREADY borne by the government
has to be considered as ?additional? when evaluating how much more new
students would cost the nation. Page 135 of the 2008 Preliminary
Educational Sector Performance Report provides the 2008 per capita costs
over a period of 5 years. Using the same rate of growth over that period
to project current government expenditure (usually called ?subvention?)
yields an average expenditure of more than $840. This gap in the analysis
alone, sadly, completely derails the CNA?s conclusions.
3. The notion that over the next two years the Ghanaian economy will
double (from about GHS45billion to more than GHS92billion) cannot be based
on any realistic projection of current growth trends. At best a 25%
expansion in GDP (as opposed to the CNA?s 100%) can be expected. Any
expanded oil revenues will be based on current planned investments in the
oil resource (i.e. we need to pre-invest in oil infrastructure for fresh
output to come on board) giving us a clear view today of what oil revenues
are likely to be over the life of a prospective NPP government, should
they win the 2012 general elections.
These grievous gaps, unfortunately, makes the CNA?s figures untenable in
the current debate.
Below, we have proceeded to use the same methodology and data sources as
they did (but only after closing the critical gaps in their analysis).
Because there is no headcount of Secondary High School (SHS) or even
Junior High School (JHS) students this year or the year before, the
analysis has to be based on projections from the available data. The
Ministry of Education (then Ministry of Science, Education, Sports &
Science ? MOESS) published the now widely referenced, ?Preliminary
Education Sector Performance Report (PESPR)? in June 2008.
In that report the Ministry provided the following enrolment figures at SHS:
2003/2004 ? 328,426
2007/2008 ? 454,681
These data points suggest a compound growth rate of 38% over 5 years. 5
years is a good time span to average out year to year fluctuations.
Using the same growth rate trends, it is safe to assume that current
enrolment, in 2012, is about:
2011/2012 ? 630,000
It is important to note that while this data includes figures for private
SHS students, it does not include information on the technical,
agricultural and vocational (TAV) sectors, where data is weakest (the 2008
PESPR report provides an estimate of about 70,000 ?trainees? in public TAV
institutions, but the sector is treated in such a peculiar fashion that it
is perhaps best to leave it out of this analysis.)
We hope to show later in our discussion that it does not really matter
which proportion of total student enrolment relates to the private sector
as opposed to the public sector. This is simply because we will be using
pre-calculated average expenditure per student, which already takes into
account the fact that the central government spends very, very, little on
students in private SHS schools.
In considering enrolment for the coming analysis, we shall be using the
transmission rates between JHS and SHS instead of gross and net enrolment
rates. The proposed NPP policy has as its core objective the extension of
?basic? education to cover SHS, and sees therefore a seamless
?transmission? of JHS students to SHS, in very much the same way that
Primary 6 pupils ?transmit? automatically to JHS today. We should, hence,
in our analysis be interested in:
1. The certain increase in the number of JHS students that shall enrol in
SHS in 2013/2014 should the NPP form the next government.
2. The fact that JHS enrolment itself is likely to expand further
considering the huge numbers of potential JHS enrolees who are currently
not in school. This shall feed into even higher enrolment at SHS in due
With that in mind, we can now examine the JHS enrolment figures provided
in the PESPR report:
2003/2004 ? 986,111
2007/2008 ? 1,224,964
Using the same 5-year compound growth rate, we can project current total
enrolment at JHS, i.e. in the 2011/2012 academic year, as follows:
2011/2012 ? 1,525,000
Consequently, at the commencement of the 2013/2014 academic year, assuming
the same rate of transmission as primary-to-JHS (67%), the total number of
enrolled students in SHS1 shall expand from the current roughly 210,000 to
[One can increase the total enrolment in SHS2 and SHS3 by the historical
trend of annual increases. Let?s however leave the figure static for
clarity of analysis.]
On the above assumption, it is estimated that upon the commencement of the
policy in 2013/2014, roughly 762,000 students shall be enrolled overall.
This is a very realistic projection.
We could however also, for future reference purposes, assume a more
conservative transmission rate, of 58%. That shall imply an SHS1 enrolment
level of 300,000 students upon the policy?s commencement, and an overall
SHS enrolment level of 720,000.
Based on identical projections, the realistic scenario suggests a total
SHS enrolment figure in year 4 of the NPP administration (year 3 in the
policy?s lifecycle) of 1.28 million.
Based on identical projections, the conservative scenario suggests a total
SHS enrolment figure in year 4 of the NPP Administration of 885,000. It is
important to point out that even ignoring the likely effects of the policy
current growth trends suggest that the SHS enrolment level in 2016/2017
shall be approximately 790,000.
For ease of analysis, we assume that in year one, all ?additional? costs
as a result of the policy shall arise because of the additional 90,000
students (in the conservative model) or 132,000 students (in the realistic
model). The other cost item, unexamined in this analysis, to keep in mind
is the ?cost of adjustment?. The cost of adjustment includes light capital
expenditure such as desks, learning materials (including text books),
public education, supervision, etc. but not heavy capital expenditure such
as classrooms, dormitories, dining halls, transport equipment, workshops
and laboratories etc.
THE ALMIGHTY EQUATION
To determine the additional (or combined marginal) cost due to new
students joining the secondary school system in year one, we employ an
elementary arithmetic relation:
Unit Current Govt Subvention (CGS) + Appropriated Household Expenditure
(AHE) = Public Sector Student Per Capita Expenditure (SPC)
That is simply to say that the cost of the current subsidy or subvention
per student at SHS plus the expenses currently borne by households but
transferrable to government in the wake of the policy equals the
additional cost per new student enrolled as a result of the policy.
Both the government?s average expenditure per student ? CGS ? and the
private expenditure per student ? AHE ? must be adjusted year by year for
inflation and currency depreciation costs in order to obtain the ?real?
amount in the future.
Based on historical trends, a conservative 15% compound annual growth rate
might suffice for direct analysis. We however chose to rely on implied
projections in this narrative, thus enabling us to relax this assumption.
Let us now examine one after the other the two inputs in the arithmetical
CGS (Unit Current Government Subvention)
According to the Ministry of Education itself:
The Per capita (the Ministry distinguishes this from the ?unit? cost)
public sector expenditure ? the ?CGS?- in 2003 was GHS191.
By 2007 the corresponding figure was GHS523.
It is instructive to note however that these figures did not include
?personnel emolument?, essentially the cost of salaries paid out to
teachers and other public sector educational workers.
The combined growth rate in the CGS government expenditure over the
time-span under consideration is thus 273%.
It can be safely projected that the corresponding 2012 figure is GHS1,452,
which is in turn equivalent to $842 in 2012 prices. It is acceptable to
use this as the initial CGS figure for the 2013/2014 commencement year.
While this figure may come across as aggressive given other estimates
provided so far in this debate, it is important to note that it is derived
by sound principles from the Ministry?s own base data, and that it does
not even include personnel emolument costs.
The terminal CGS (government average expenditure per student in 2016/2017)
can be derived by projecting the 2008 figure at 2012 prices yielding
$1,220. It is interesting to point out that at 3% per annum the dollar
inflation rate is not completely negligible but for clarity of analysis we
will leave it static.
We now need to compute the AHE.
AHE (Average Household Expenditure)
We use the GLSS-5 data (notwithstanding the widely known concerns about
occasional granularity issues at district level).
The average secondary school expenditure per student per household is
provided in the September 2008 GLSS-5 report as follows: GHS244, which
corresponds to $256 (at 2006 dollar prices of 0.95 GHS).
If we were to follow the CNA?s lead and assume that government shall
assume 85% of the costs involved, then the adjusted AHE figure for this
analysis comes to $217. Given the breakdown of costs in the GLSS-5
however, there is no reason why government cannot choose to absorb the
entire cost if the policy is to boost enrolment. Since boarding students
outnumber day students, the transport component of the AHE build-up, the
one cost item easily discounted, is very small.
SPC (Total Policy-Related Cost Burden Per Student)
From an abundance of caution, we have decided to use the 2008 CGS level
throughout the remaining analysis, thus producing ultra-conservative
In that light the cost per each new student is $1100 (CGS + AHE). In the
scenario where government chooses not to assume the full complement of
direct household costs the corresponding figure is: $935.
TOTAL POLICY-RELATED COST BURDEN ? CONSERVATIVE
In the ?conservative scenario?, this amounts to $134 million or $114
million, should government assume only partial responsibility for private
costs. Note however that this figure relates strictly to additional
enrolment ?following the introduction of the policy?. The government must
also absorb the household expenditure of those SHS1 students who ?would
were enrolling anyway?. This is an additional $54 million, bringing the
total cost to $184 million or $168 million in the conservative model.
In the ?realistic scenario?, this amounts to $154.5 million or $133
million plus $54 million, bringing the total cost to $208.5 million or
TOTAL POLICY-RELATED COST BURDEN ? REALISTIC
At this point, the analysis is more straightforward because we shall be
able to use total SHS enrolment to compute the total cost burden.
Based on assumptions already discussed, total JHS enrolment shall stand at
1.906 million students.
In the conservative model, total enrolment at SHS level shall come to 1.1
million. Total expenditure shall therefore come to $1.216 billion or
$1.0285 billion (in the scenario where government assumes partial
responsibility for private costs).
In the realistic scenario, total enrolment at SHS level shall stand at
1.277 million. Total expenditure shall therefore come to $1.404 billion or
The median projection is thus: $1.340 billion or $1.1115 billion.
These calculations do not explicitly accommodate the policy?s adjustment
and administration costs, defined above to include learning materials,
furniture, public education etc. There has furthermore been no accounting
for capital expenditure such as buildings and transport equipment etc. No
personnel emolument or training costs are included in the CGS calculations
providing additional latitude for any downward adjustments due to error.
Because the data used for the CGS calculations are already averaged out,
there is no need for specific disaggregation of rural ? urban, private ?
public, or day ? boarding sub-data.
Clearly, the use of GLSS-5 and Ministry of Education data, and the narrow
focus on SHS intake analysis in no way revises the earlier estimates
downwards. If anything at all, the higher averages, due to higher
resolution data, merely exacerbate the cost implications of the fee-free
In future commentary, we shall investigate the capital costs of
adjustment, before moving on to examine the cost-benefit matrices and
overall structural and other objective social and economic merits of the
IMANI Center for Policy & Education (www.imanighana.org); syndicated by