Introduction
India and Pakistan are two of the largest spenders on defence in the world. In spite of the ongoing (2003) attempts to give a new thrust to normalisation of bilateral relations, the two countries continue with their major drive at militarisation. The current phase of military expansion is, in part, linked to the decision by both countries to become nuclear powers. It is also a continuation of the trend established in the 1990s when India began a modernisation programme and Pakistan, in response, hoped to keep pace with its neighbour. Both programmes have imposed huge financial and opportunity costs on the two economies. With India and Pakistan deciding to continue on the nuclear path, the cost of nuclear weaponisation is only going to increase in the years ahead. In short, militarisation in the two South Asian neighbours, which has always imposed costs on their development efforts, will continue to do so in the future.

Both countries figure among the biggest spenders in the world. According to the most recent data compiled by the Stockholm International Peace Research Institute (SIPRI), India's military defence expenditure, when measured on market exchange rate, amounted to US$. 12.9 billion in 2002, making it the country with the 11th highest military expenditure in the world. Pakistan does not figure in SIPRI's list of the 15 largest spenders in the world. However, when national military expenditure is compared on a purchasing power parity (PPP) basis, the picture is quite different. According to SIPRI, India with an expenditure of US$. 66.2 billion in 2002 ranks third in the world. Only the U.S. and China made a larger outlay than India on defence last year. Pakistan with a military outlay in 2002 of US$. 14.2 billion (PPP terms) ranked 15th in the world.

An attempt at assessing the impact of military expenditure on the economies of India and Pakistan has to begin with an estimation of the true size of the burden of defence. Unfortunately, in India and Pakistan, the budgetary figures on defence do not give a complete picture of total outlays. The data for India is far more transparent than for Pakistan (where there is only a single line figure on defence in the official documents) but the problem in both countries is that budgetary figures under-estimate the true burden of militarisation.

SIPRI estimates cover a reasonably broad definition of what constitutes defence expenditure, but in as much as these estimates too rely on published documents, the assumption must be that the SIPRI data is also not complete. Presented below are SIPRI estimates since 1990 of the defence expenditure as a percentage of the gross domestic product (GDP) for both India and Pakistan as reproduced in World Bank data.

TABLE 1: Defence Expendture as % of GDP
Year	India	Pakistan
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001	2.5 2.3 2.4 2.3 2.2 2.1 2.2 2.2 2.3 2.3 2.5	5.8 6.1 5.7 5.3 5.3 5.1 4.9 4.8 4.6 4.4 4.5
Source: World Bank Indicators (World Bank Data Base: 2003)

Pakistan has consistently shown a higher level of military expenditure-GDP than India. However, before analysing the data in Table 1, a comparison of defence outlay-GDP proportion of a few countries is given below:

TABLE 2: Defence Expenditure as % of GDP
China India U.S. Russia Pakistan Turkey Israel S. Arabia	2.1 2.3 3 3.8 4.6 5 8.1 13.2
Source: As in Table 1

It appears that in larger countries, defence expenditure as a proportion of GDP is generally lower than in the smaller countries. Thus, China, India, the U.S. and even Russia have lower defence-GDP levels than Pakistan, Israel, Turkey and Saudi Arabia, although it is difficult to assert that the first group of countries shows a significantly lower level of militarisation than the second group. One can speculate that there is always a 'minimum' level of military infrastructure that all countries have to establish, which is reflected in smaller countries showing a higher defence-GDP ratio. However, even if this is true, one cannot deny that a high degree of militarisation is responsible for the astronomically high defence-GDP estimates for Israel and Saudi Arabia.

To return to Table 1, India's defence expenditure in relative terms was high at the beginning of the 1990s (it reached a peak of 2.8-2.9 percent of GDP in 1989), before gradually declining and picking up from the mid-1990s onwards. For Pakistan, a somewhat similar pattern holds, except that defence expenditure does not pick up in the mid-late 1990s as it does for India. This does not mean India and Pakistan have been giving increasingly less importance to military spending. Reasonable explanations can be offered for each case. However, one must first reiterate that Pakistan has always had a higher level of defence-GDP ratio than India, though Pakistan's higher level of defence-GDP outlay could be explained partly in terms of 'minimum outlay' hypothesis.

Defence Spending During the 1990s
The defence spending-GDP ratio is the standard yardstick of measurement of the burden of defence. As will be argued later, this is not the best indicator of spending. But what of spending in absolute terms? SIPRI estimates of military outlays in the two South Asian countries during the 1990s provide a basis for comparison. This is presented in Table 3

TABLE 3: Military Expenditure (in US$ million, at 2000 prices)
Year	India	Pakistan
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002	8051 7532 7209 8137 8109 8340 8565 9307 9387 10482 10900 11837 12882	2636 2823 2997 2993 2917 2965 2961 2837 2833 2858 2867 3071 3176
Source: SIPRI estimates (2003)

The facts show that in India expenditure in real terms initially declined during the 1990s only to pick up gradually from the mid-1990s. From the late 1990s onwards there has been a sharp acceleration. Indeed, between 1998 and 2002 an interval of just four year -- there was a 37 percent increase in outlays. This is a huge increase in real terms. The rise would have been even larger but for the fact that in recent years capital spending has turned out to be less than budgeted for.

The trend is different for Pakistan. After an initial spurt, military outlays were more or less stagnant in the first half of the 1990s, before falling slightly and showing a moderate increase in the initial years of the first decade of the 21st century. Although Pakistan shows a smaller increase than India, the country remains, as argued below, more militarised than India.

Defence Spending: (1961-2003)
Graph I presents the trend in the long-term movement of the defence spending-GDP ratio in India and Pakistan from the 1960s onwards. The chart is based on a compilation of estimates from different sources and must, therefore, be seen as presenting only a broad picture of the trend over the last 40 years.
The graph shows some interesting and at the same time well-known trends. First, Pakistan's spending has always been higher than that of India. Two, there have been spikes during periods of wars between the two countries as well at times of tension with other countries. For India, the spikes have been during the early 1960s (after the war with China), mid-1960s and early 1970s (wars with Pakistan) and during the late 1980s (modernisation and large arms imports). For Pakistan, the spikes have been other than during wars with India during the 1980s when Pakistan was involved in the Afghanistan war.

The decline during the 1990s has taken place in a specific context. In India, the conditions of a structural adjustment programme with the IMF in the early 1990s meant that defence expenditure was capped, before it could gradually be increased from the late 1990s onwards. In Pakistan, the defence-GDP ratio continued to remain steady until the late 1990s, when the country went in for an IMF loan which came with similar conditions on defence spending.

The Indian Case
An attempt has been made here to estimate the full extent of defence expenditure for India in recent years. Where data presented earlier have been either of official nature or of SIPRI estimates, the exercise here takes a larger sweep of all defence, para-military and related expenditure an aspect which is of particular importance to the nuclearisation of the two countries.

Indian official estimates of defence spending, as reported in the budget, cover current and capital expenditure in the three forces and also research and development. They do not cover (i) pensions, (ii) para-military expenditure (iii) defence-related atomic energy outlays and (iv) defence-related space outlays. SIPRI estimates include the first and second items of outlays.

In this paper, a broad coverage of defence expenditure in India has been covered. This coverage is necessarily approximate and is based on a number of assumptions. The source, however, remains the budget documents of the Government of India. Para-military expenditure is as reported in the Home Ministry outlays. The biggest assumption is regarding defence-related space and atomic energy expenditure. It is assumed that 25 percent of all outlays on space and atomic energy (other than for nuclear power generation) have a defence orientation. This is a reasonable assumption to make, given that much of India's ongoing nuclear programme will be based on work done at the nuclear and space research establishments.

TABLE 4: 'Official' versus 'Alternate' Estimates for India (Rs crores, at current prices)
Year	Official	Alternate
1995-96 1996-97 1997-98 1998-99 1999-2000 2000-01 2001-02 2002-03 2003-04	26879 29498 36099 41200 48504 54461 57000 56000 65000	33940 37336 46120 54255 66232 72308 75170 73777 83955
Source: Computed from Government of India Budget documents

In 2003-04 (budget estimates), the budgeted outlay, according to the alternative estimate of defence expenditure is 29 percent more than the official figure. The gap between the official and alternate estimate widens in the late 1990s, exactly when Indian defence expenditure begins to accelerate. The higher outlays, according to the alternative estimate, take the defence spending-GDP ratio to much higher levels. Thus, while the official figures suggest a defence spending-GDP level of 2.5 percent in 2001-02, the alternate estimate leads to a ratio of as much as 3.3 percent. Clearly, when all aspects of defence spending are taken into account, outlays are much higher than the figures that form the basis for the official data.

It is more than a reasonable speculation that a similar exercise conducted for Pakistan
will show a similar (perhaps even larger) gap between the official and true levels of defence spending. Unfortunately, the same exercise cannot be done for Pakistan because of the much greater opaqueness of official statistics in Pakistan.

Real Burden
The analysis has so far focussed on defence expenditure as reflected in the spending-GDP ratio. This, however, is not the best measure by which to assess the burden of the defence sector on the economy. Since military expenditure is incurred entirely by the government and as governments in developing countries also have to carry out important functions in the social sector and infrastructure, what does matter ultimately is the demands the defence sector places on the resources of governments.

Military spending also has an impact on the private and non-government sectors. However, the first charge is on government, specifically central government spending. Here the data are revealing. First, the size of military expenditure at the central/federal government level is huge. According to comparable World Bank data, military spending as a percentage of total central government expenditure was 14 percent in India in 2001. The figure for Pakistan for the same year was much higher at 23 percent. In both India and Pakistan, defence is the second largest item in central/federal government spending. Indeed, if one were to exclude interest payments then defence (capital and revenue) expenditure is the largest item of expenditure. Moreover, historically the burden of the military on central government expenditure in Pakistan has been much higher than in India. Such a high proportion of government resources being consumed by the military sector does inevitably have an impact on government outlays in the social sector.

The second and related point is that military spending dwarfs government spending on the two main social sectors of education and health. This is reflected in the following World Bank data for 1999 the latest year for which information is available. The comparable data are for public spending as a proportion of GDP, not of total government expenditure, but the data still tell the same story.

TABLE 5: DEFENCE VS EDUCATION AND HEALTH (% of GDP, 1999)
Country	Public Spending on Health	Public Spending on Education	Defence Spending
India	0.91	4.1	2.3
pakistan	0.9	1.8	4.6
Source: World Development Indicators Database (World Bank, 2003)

Defence expenditure dwarfs public spending on health in both India and Pakistan and it is far above public spending on education in Pakistan. In India, public spending on education in 1999 was considerably more than on defence. The misplaced priorities are most evident in Pakistan, which in relative terms spends far more on defence than India but also spends less on the social sectors than India. The comparison with respect to central government expenditure is actually worse than presented in Table 5. 'Public spending' refers to intervention by both central/federal and state/provincial governments and in the case of education refers to outlays at all levels. If one were to compare only central government spending on defence with that on the social sectors, then the situation is far more stark.

The picture is most graphically presented in the case of India in Graph II which plots the movement of military spending against central government spending on elementary education, education and health. Central government spending on defence in India is manifold that on education and health. As the well-known economists Amartya Sen and Jean Dreze have pointed out, India's defence spending is more than three times the combined central and state government expenditure on health.

South Asia has some of the worst indicators in health and education, yet the governments of India and Pakistan clearly prefer to spend more on defence.

Comparison with the World
Both India and Pakistan are allocating a considerable amount of resources to the defence sector. Arms imports as a proportion of total imports are higher in Pakistan than in India. According to World Bank data, Pakistan's arms imports constituted 9.7 percent of total imports in 1999 (the latest year for which data is available), compared to 1.6 percent for India in the same year. This ratio varies in both countries from year to year, but in general the ratio for Pakistan is around 6 percent while that for India it is 2 percent. Clearly, the imports of arms place a bigger strain on Pakistan's balance of payments than on India.

Another indicator of the greater role of defence in Pakistan's economy is in the size of the military force. India of course has the larger army, with 1.3 million personnel in uniform, while Pakistan has less than half this number, 590,000 (1999, World Bank data). However, in relative terms, the Pakistani army is proportionately bigger. The Indian defence personnel population constitutes just 0.30 percent of the labour force. The Pakistani military strength, on the other hand, is as much as 1.2 percent of the labour force - four times as large as in India.

With respect to averages of low-income countries and of the world as a whole, India and Pakistan allocate considerably more resources to defence. Table 6 presents statistics for India, Pakistan, averages for all low-income countries and for the world.

TABLE 6: INDIA, PAKISTAN AND THE REST OF THE WORLD
Category	year	India	pakistan	Low-income countries	World average
Defence Spending-GDP (%) Defence Spending-Central Govt. Expenditure (%) Military Personnel-Labour Force (%) Arms Imports-Total Imports (%)	2001 2001 1999 1999	2.5 14 0.3 1.6	4.5 23 1.18 9.7	2.3 13.1 0.56 2.1	2.3 9.8 0.7 0.9
Source: World Development Indicators Database (World Bank, 2003)

It is clear that India and Pakistan are showing higher levels of defence spending than other low-income countries or the world as a whole. Where the size of the military population is considered, the Indian figure is lower than the global average while for Pakistan it is larger. In arms imports, Pakistan spends relatively much more than what the low-income countries do. On the whole, the message that comes through from these figures is that India and Pakistan give considerably more importance to defence than other countries that are at roughly the same level of development.

Guns versus Butter?
The argument presented here suggests that money allocated to defence is a diversion of scarce resources from other sectors; and that for countries like India and Pakistan, which still belong to the group of low-income countries, this is a diversion they can scarcely afford. This is what economists refer to as the 'Guns versus Butter' argument the choice is between more arms and more butter. More of one means less for the other.

In theory, this argument is valid only when a country is at a stage of what is called the 'production possibility frontier' -- resources are fully deployed, and only a re-allocation between competing sectors is possible. One criticism of the application of the guns versus butter argument to developing countries is that these countries are not at the production possibility frontier, therefore it is not a question of guns versus butter in allocation of resources. It is possible, the counter-argument would imply, to increase resource allocation to defence without cutting down on investment in important economic and social sectors. Another criticism of the guns versus butter argument is that if India or Pakistan do reduce defence spending, government spending on primary education, health care or any of the other important areas, in the countries will not necessarily follow.

Both counter arguments are valid, but neither can be offered in defence of the high level of military spending the two countries have been incurring and the future levels they have committed themselves to.

The most powerful critique of an expansionary military spending that has been made in recent times is the one offered in 2002 by the economists, Jean Dreze and Amartya Sen. The critique is of India's policy since the late 1990s and covers India's nuclear weapons policy as well. But the argument is equally valid for Pakistan. The argument made by the two economists is that there are many 'social costs of militarism' of the kind pursued by India. One, rising military expenditure imposes substantial opportunity costs on government priorities like health care and primary education, even if every rupee saved in defence does not lead to a corresponding hike in social sector spending.

Two, nuclear weaponisation leads to increased insecurity in South Asia. Three, nuclear weaponisation will lead to an arms escalation in South Asia, which will end up in further diversion of scarce resources to the defence sector. Four, there is not merely a diversion of economic resources when countries like India and Pakistan embark on an arms race. There are also the demands made on 'the time and energy' of political leaders, government officials and the public at large. Fifth, military expansionism leads to a diversion of scientific and technological resources to the defence sector. The research and development expenditure in defence, space and nuclear field in India constitutes over 60 percent of total government research outlay in the country.

All told, there is indeed a trade-off between guns and butter, '...there is much evidence that military expenditure adversely affects economic performance by 'crowding out' other uses of scarce resources, such as private investment and social spending. In the case of India, recent increases in military expenditure are bound to affect prospects for a much-needed expansion of public expenditure on health, education, social security and related matters. Indeed, given that the bulk of money is pre-committed in the form of public sector salaries and interest payments, mobilising additional resources for the social sectors is a major challenge. Restraining military expenditure is among the few available options (others include wasteful subsidies and expanding the tax base), and in that sense the trade-off between military and social expenditure is quite sharp.(emphasis added)

Schools for Children or Nuclear Weapons?
The discussion above has made no reference to the decision by India and Pakistan to go nuclear. Nuclear weapons introduce a qualitatively new dimension to defence spending in the subcontinent.

If the rising burden of military expenditure in the late 1990s in India is not cause enough for concern, we now have the decision to go in for open nuclear weaponisation to contend with. Nuclear weaponisation will not just be costly, it also threatens to change the nature of the state and engage India and Pakistan in a dangerous race of nuclear proliferation. As far as expenditure is concerned, there is very little evidence to suggest that nuclear weaponisation will lead to savings in conventional arms outlays.

There is very little information on the likely costs of a nuclear weapons programme in Pakistan, but there is enough data available in the public domain in India to make an informed estimate about what an Indian weapons programme could cost. This could serve as a benchmark for the Pakistani programme.

It is often argued that a nuclear weapons programme for India will not be expensive. The reasoning is that India already has a nuclear infrastructure in place and that the additional expenditure required will not be much. Another reason given is that India will not duplicate the gigantic nuclear weapon models of the U.S. and the former Soviet Union and instead develop a small nuclear arsenal. This, however, is not correct reasoning.

First, as events since 1998 have shown, nuclear arms have not reduced spending on conventional arms in India. If anything, the increased insecurity that nuclear weapons have brought to the region has led to higher conventional arms spending. Second, while in theory one can make the case for a small nuclear weapons arsenal, in practice the demands for an expanding arsenal will keep growing. This has already been taking place with signs of an inter-service rivalry in India, major import/joint development programmes for supporting infrastructure (command and control) and expanded/new programmes like the anti-missile defence systems etc. Much of this is reflected in the numerous deals and proposals India has been exploring with Israel, Russia and even the U.S. All this suggests that the elements of a new arms race are in the making in the subcontinent.

A very conservative estimate of the cost of an Indian nuclear weapons programme suggests that at a minimum this would costs Rs. 800 billion over a decade at 1998-99 prices, or Rs. 700-800 billion a year. This is equivalent to an incremental cost of 0.5 percent of India's GDP every year. The dollar costs over a decade on an Indian nuclear weaponisation programme will be around US$. 16-19 billion (at the average 1998-99 market exchange rates) or US$. 81-93 billion (at the 1999 purchasing power parity, PPP, exchange rate). The larger component in these costs would be the outlays on delivery systems (missiles and nuclear submarines) and on a command and control system.

To give an idea of the financial implications of a Rs. 700-800 billion Indian nuclear weaponisation programme spread over a decade:

An Indian nuclear weaponisation programme that would cost 0.5 percent of GDP a year is equivalent to the annual cost of introducing universal elementary education in India.. This 'high' cost was for years cited as one of the reasons for not universalising elementary education in India. The question then is of choosing between sending every Indian child to school and acquiring nuclear weapons, both of which are going to make similar financial demands on the Government of India. Although India's Parliament in 2001 enacted an amendment to the Constitution guaranteeing elementary education to every Indian child, the initial financial allocations suggest that the government is giving a greater importance to nuclear weapons than to universal elementary education.

In recent years, India has entered into a number of arms import agreements and is planning to purchase more from the U.S., Russia, Israel, the U.K. and France. These deals are not for nuclear weapons per se, but they are for a number of weapons and defence systems that will become an integral part of a command and control infrastructure that India is building for its nuclear arsenal. They include anti-missile systems, intelligence radar, delivery aircraft and leasing of nuclear submarines. In addition there has been a new thrust to domestic research and development directed towards missiles. If even some of these costs are allocated to the Indian nuclear weapons programme as they must be then the annual cost of the Indian weapons programme will end up as much more than 0.5 percent of GDP.

In all this, Pakistan has not been found wanting.