The author is a professor at the University of the Philippines School of Economics. This article was originally the author’s presentation at the Annual Meeting of the Philippine Economic Society last April 10 at the Bangko Sentral ng Pilipinas in Manila City.

Persistent budget deficits traceable to a falling tax effort since 1998
have focused attention on the need to reform tax administration in the
Philippines. At the moment, there are a number of proposals to reform
tax administration. These include House Bills 5054; 5465, 5546, and
Senate Bill 2463.

The template for most of these bills is from the original bill HB 5054,
also known as the “IRMA bill” (since it sought to create an “Internal
Revenue Management Authority” while abolishing the Bureau of Internal
Revenue). This has since been modified, most preeminently by HB 5465 or
the “NARA” proposal (standing for the “National Authority for Revenue

These proposals may be collectively called as BIR-reform bills. While
they might vary in some detail, all of them make the radical proposal
of abolishing the existing Bureau of Internal Revenue and replacing it
with a new authority vested with corporate-like powers. This authority
is to be headed by a chief executive appointed by a board composed of
both ex-officio government and full-time private representatives.

In this column, I aim to present the economic rationale for these
reforms in the light of results in the theory of incentives and
contracts. In doing so, we hope to highlight what is essential,
incidental, or contrary in the proposed measures and hopefully serve as
guide to public debate on the issue.

Weberian bureaucrats and efficiency wages

In a good deal of public discussion of the IRMA, NARA, or its various
incarnations, what has been emphasized has been the adoption of the
corporate form, and some have chosen to describe these reforms simply
as a “corporatization” of the BIR. For would-be detractors, this is a
smart tack, considering the poor record of public corporations in the

Such an emphasis on the “corporatization” of the revenue-collecting agency is entirely misplaced, however.
The real innovation contained in the reform bills is not with respect
to the adoption of the corporate form, but the change in the internal
structure of incentives in that agency. Put another way, if it were
only possible to change the internal structure of incentives without
adopting a corporate form and remaining within the government
structure, then one would not have to go through the pain.

Like many other agencies in the government, however, the current BIR is
structured to approximate the ideal of a Weberian bureaucracy. The
present system presumes that tax collection work is routine, much like
sorting letters at the post office. For such tasks, “low-powered
incentives” are sufficient.

Moderate but stable pay plus secure tenure are usually sufficient to
elicit a minimum, standard effort from a Weberian bureaucrat.

While separation from such a system is still possible when performance
becomes egregiously substandard, the threshold of performance is, in
fact, not particularly high. On its own terms, such a system is
probably sufficient when output, which often consists of providing
public goods, is typically difficult to observe.
Higher-powered incentive schemes would be difficult to design in such circumstances.

Essentially, we can think of the Weberian bureaucrat as being kept in
line through an efficiency wage, which requires that the “gain” from
malfeasance should be less than the “rent” that would be lost from
separation from the service. The latter is the difference between the
actual and opportunity wage, multiplied by the probability of detection
and the long-term value of the relationship.

g < p(w – wo)N

Hence the minimum efficiency wage required to keep a bureaucrat honest is

w = (wo + g/Np)

If this efficiency wage is higher, the potential gain from malfeasance
is higher while the risk of detection and long-term prospect of the
employment are lower. The stability of employment due to permanent
tenure substitutes – to some extent – for a higher wage (N>1 is
presumably at a maximum). The nature of routine public-service tasks
makes g or the “gain” small and also makes malfeasance easy to detect,
i.e. p large.
Such a system has not worked well for the Philippines and particularly
for an agency such as the BIR for several reasons. First of all, for
many positions of responsibility – particularly such as those in the
BIR – “gain” tends to be very large. Second, the probability of
detection p is also very low, partly because of the wide latitude for
executive discretion as well as the unreliability of conviction under
the justice system.

These factors imply an efficiency wage well beyond what the government
can typically afford. If, for example, a bureaucrat’s potential gain
from corruption is as low as P1 million annually, with a probability of
detection of as high as 50% (even assuming N = 10), that bureaucrat’s
salary would have to be P200,000 more than any outside wage to keep him
from straying. This, to put it mildly, is more than anything the Salary
Standardization Law would allow.

Finally something should also be said about wo, the opportunity wage.
In some social contexts, the minimal efficiency wage for bureaucrats
need not be very high, since a malfeasant’s post-service pay would be
very low indeed. In such instances, few others would hire someone who
was dismissed for cause from the service. Here, social customs and
practices would be relevant.

In closely knit communities where reputations and one’s “good name”
might matter, an offender might be socially shunned or ostracized. This
is not currently so for Philippine society where the operative motto
seems to be pecunia non olet, and where wealth regardless of its
provenance can always buy a reputation – sometimes even seats in the

There are obvious difficulties, therefore, why the current system of
bureaucratic rewards in the Philippines has not worked well.

A particularly tight feedback has to do with tax collections: Because
budgets are low to begin with, and therefore the current public wage is
below the efficiency wage. The result is that malfeasance increases,
causing tax collections to fall even further, making it even more
unlikely that the efficiency wage will ever be met.
As against efficiency wages, the reform bills emphasize the central
importance of introducing incentive contracts as a paradigm to improve
the revenue effort. Under an incentive contract, the larger part of pay
is not a fixed amount but is dependent on productivity or effort,
however this is measured. Ideally, high pay or reward is associated
with high effort or the observable results of that effort. In a real
sense, even with the same job, employees “choose” their compensation by
selecting their level of effort. Probably the most familiar example of
such contracts is the piece-rate system in many factories. At the level
of executives in the private sector, the implementation of performance
bonuses and stock options is also well known.

What is somewhat novel is the intent to use such contracts in a
government agency discharging public functions. The Internal Revenue
Management Authority (IRMA) and National Authory for Revenue
Administration (NARA) bills propose that the chief executive officer
(CEO) of the new agency be governed and bound by a “performance
contract” with the revenue board. Here, the CEO will sign on to a
contract that requires him to raise collections to a certain absolute
amount. The contract might provide that any overperformance be rewarded
by an executive bonus, while underperformance might lead to a pay cut.
In the worst case, consistent underperformance could lead to
replacement, as it does in the private sector.

The basic shortcoming of a bureaucratic organization based entirely on
fixed pay is that it is too blunt to encourage performance where
variable effort and discretion are crucial for results. Unlike
bureaucratic wage contracts, incentive deals recognize the wide
latitude that revenue executives possess in determining the scope and
intensity of enforcement of tax laws.

Revenue collection is not a task bureaucrats discharge mechanically as
it entails a good amount of discretion. The approach of
bureaucratic-wage systems is progressively to delimit the scope for
such discretion by putting in place rules and complicated monitoring
mechanisms (e.g., civil service and auditing rules). By tying rewards
to performance, incentive contracts minimize external restraints and
explicitly recognize the leeway executives have. These also make
employees realize there is a choice between high and low effort.

Unlike other routine bureaucratic tasks, tax collection is one that
lends itself to some measurement. This is important since an incentive
contract can be designed only if some objective measurement of output
At the national level the “tax effort,” properly adjusted for income
growth and any new tax legislation, is literally a measure of
performance. In effect, incentive contracts abandon the mode of
monitoring that wage contracts imply.

Under a system of low-powered bureaucratic pay, no serious consequences
arise unless effort or confidence falls drastically below a certain
level; that threshold itself is defined by the standard for what is
required for dismissal from the service and lengthy court battles.

With high-powered incentives, both reward and tenure are more finely
tuned. It should be possible under such a system to replace a CEO or a
revenue district officer not necessarily because they committed an act
for which they can be charged before the Ombudsman, but simply because
they failed to fulfill a mutually agreed collection target.

To use an analogy, a bureaucratic wage contract are like ordinary
switches or circuit-breakers. Incentive contracts, meanwhile, are like
switches that include light-dimmers, or thermostats.

One thing the current bills do not do is to determine exactly who are
to be covered by incentive contracts and who are to continue to be
governed by efficiency wages. This is something that must be spelled
out in the implementing rules and regulations or by the proposed board
of the new authority.

It is obvious that such high-powered incentives would be most relevant
and applicable to the highest echelons, especially the position of CEO,
where leeway for discretion is high and there is some objective measure
of results.

However, such arrangements are likely to be impracticable for positions
involving routine jobs and those where the evaluation of a person’s
effort is more likely to be subjective. Exactly where the line must be
drawn between these two is something that still needs to be determined.
The bills now filed contemplate this issue being resolved by leaving it
up to the CEO to “establish a performance-based management system which
shall govern the selection, hiring, appointment, transfer, promotion,
or dismissal of all personnel”.

Accepting the principle of incentive or performance contracts at the
highest levels requires one to consider other incentive-design features
in the matter of hiring and firing. Incentive contracts are premised on
sufficient latitude of action for executives, and the latter would not
sign on to high-powered contracts if they did have leeway to dispose of
personnel and resources as they saw fit in order to achieve their
objectives. It is for this reason that the reform bills give a good
deal of authority to the CEO in both the appointment and removal of
personnel. Hence not only is the CEO given the authority to design the
system, but also to actually exercise the power to “appoint, fix the
remuneration and other emoluments, and remove personnel of the
It is in the transition between the present Bureau of Internal Revenue
(BIR) and the new authority that the various versions of the reform

The original Internal Revenue Management Authority (IRMA) states clearly that:

” No preferential or priority right shall be given to or enjoyed by any
personnel for appointment to any position in the new staffing pattern,
nor shall any personnel be considered as having prior or vested rights
with respect to retention in the Authority or in any position which may
be created in the new staffing pattern, even if such personnel should
be the incumbent of a similar position prior in organization.”

However, the National Revenue Authority (NRA) bill reverses this explicitly:

” Personnel of the (BIR) at the time of the approval of this Act who
apply for positions in the Authority shall enjoy preferential
absorption into the new organization. For the purposes of this
provision, “preferential absorption” means that (BIR) personnel who
meet the new qualification standards and selection criteria shall enjoy
priority in the hiring of personnel for the Authority. Those who cannot
be absorbed due to either the inability to meet the qualification
standards and selection criteria or the nonexistence of a suitable
position in the Authority shall be deemed separated from the service.”

. The difference – which was seen as sufficient to merit the filing of
a new bill – was explicitly an accommodation to “a reaction from
employees of the BIR who expectedly complained against the possibility
of losing their jobs.”
It has been explained that the logic of incentive contracts entails a
wide latitude being given to executives. In an ideal situation, a chief
executive officer (CEO) working under an incentive contract ought to be
allowed to bring in a team on which he or she can rely. The NRA
provision works against this logic, however, by saddling the new
authority and its CEO with the BIR’s previous personnel. There are a
number of risks such a move presents to the reform scheme.

First, such a move would have an external effect on the very hiring of
the CEO. Potential CEOs may balk at having to work with key staff not
of their own choosing. Hence the government may be faced either with a
shortage of applicants for the post. Alternatively, with the new agency
advertised as being populated by former insiders in the old agency,
there is a great possibility that the only ones willing to take on the
CEO job are themselves either unsuitable or non-serious (adverse

Related to the foregoing, a second risk are disputes and distortions
that may arise in the incentive contracts the CEOs may write with the
board. Since the CEO will not have been allowed the initial latitude to
select his or her own staff, this is a fact that can be brought out in
subsequent evaluations of performance. Then the observed nonattainment
of performance targets would be equivocal and might arguably be
attributed – not to the CEO’s failure – but to the lack of cooperation
from former insiders.

Subsequent finger pointing between the CEO and the staff not of his
choosing would then be more difficult to avoid, and the power of the
incentive contract binding the managers would be reduced. Obviously
such a risk would be averted if, to begin with, the CEO were allowed a
free hand on selecting key personnel.

These risks would not be alleviated, even by the proviso that the
absorption of old personnel should be prefaced by “qualification
standards” and “selection criteria,” or, as some have proposed, even
“lifestyle checks” on former BIR employees.

Especially in the recruitment of people for positions of trust, uniform
standards and qualifications are typically crude and inadequate; CEOs
will typically supplement these with their own subjective evaluations,
some of which are actually noncomparable and nonquantifiable. The logic
of incentive contracts is to allow executives to exercise their
discretion fully in these matters, since it is they who will ultimately
reap the consequences, for good or ill. By contrast, Sec. 20(a) of the
NRA bill harks back to the method of setting bureaucratic norms for
recruitment; it restricts the CEO’s leeway in recruitment and
effectively replaces this with a rigid rule for seniority. It is in
this sense that the NRA bill provisions on “preferential absorption”
run against the logic of the very bill itself.

There may well be good reasons that a CEO would wish to rehire a significant number of personnel from the present BIR.

Foremost among these is the fact that many of them will have a great
deal of experience. But the importance of experience as against other
qualities of staff should, properly speaking, be regarded as a call by
management, not an imposition by the law.

I have approached this issue primarily from the viewpoint of
safeguarding the integrity and efficiency of future contracts under the
proposed new revenue authority.

The proponents of the NRA bill in its relevant provisions, on the other
hand, seem concerned with extending the integrity of contracts under
the existing BIR. This appears to be unnecessary.

In the first place, legal provisions and civil service rules clearly
apply regarding the fate of positions in agencies that have been
abolished. Secondly, the law also provides for the compensation of
those who have been terminated through separation or retirement
benefits, about which more below.

If any rule on “preferential absorption” is instituted at all, it
should apply only to positions that will not be covered by incentive
contracts in the authority to be established. This would exclude most
positions in top to middle management.

The reasoning from the incentives viewpoint is clear: those to be
compensated with efficiency wages are presumably those who exercise
little or no discretion in the overall affairs of the authority. They
can be granted the usual wage- and tenure-security packages under
existing labor laws. Standard qualifying procedures may be properly
applied to recruiting these, since they are destined to perform routine
tasks whose outcomes are easily monitored.

By contrast, positions of trust and confidence should be covered by
incentive contracts, with the implied absence of any claim to tenure.
Indeed a good number of such positions should be coterminous with that
of the CEO.

A word must be said finally about an apparently innocent issue, which
is the amount of severance pay to be paid. The IRMA bill (Section 21)
states that:

” The officials and employees of the former Bureau of Internal Revenue
who are separated from the service as a result of the organization of
the Authority shall be allowed to retire or be separated from the
service under existing laws.”

Existing civil service rules allow for one-month’s pay for every year
of service rendered. As a concession to the protests of the BIR
employees. The NRA bill, on the other hand, promises separation

” which shall be in addition to all gratuities and benefits to which they may be entitled under existing laws”.

The NRA bill states that it intended to correct the “mistaken notion”
that there would be no retirement or separation benefits. It went
further however, by holding out the possibility of separation benefits
over and above what existing laws allow. The greater generosity of the
NRA version was also motivated by the need to appease the current
employees. Some in Congress are now said to offer three- to five-months
pay per year of service to those to be separated. Needless to say, such
amounts would be budget-busters at just a time the country least needs

Proponents of the NRA bill were obviously motivated by some form of the
Pareto Compensation Principle: if the gains from the policy measure are
so great, they could compensate the potential losers from it.

One problem with this application, however, is the precise
identification of “losers” in this context. While one can readily go
along with the proposition that “not everyone in government is
corrupt,” this is far from saying that “no one in government is corrupt
– particularly in the BIR.”

Extraordinary compensation may go some way towards justly assuaging
those genuinely harmed by retrenchment. But the law is carrying a leaky
bucket when it promises generosity to all, even to those who profited
from their positions and amassed wealth. Viewed this way, one is no
longer certain where Pareto Compensation ends and where coercion