What is policy monitoring? Why should agricultural polices in developing countries be subject to monitoring? What aspects of those policies should be monitored? What indicators are available to be used? And how should one proceed to develop those indicators in a constructive and transparent manner? These questions are central to the theme of this paper, and so some basic definitions and propositions are useful at the outset.
In this paper we define the process of monitoring a policy or set of policies as assembling quantitative (and in some cases qualitative) indicators that can be traced over time and compared across countries. The indicators themselves are inputs into the distinct processes of quantitative modelling of market and trade effects of policies and the assessment of how well the policies are achieving their objectives. Thus for monitoring to be useful it should provide a framework for the regular collection of data and have as its output indicators that can be used directly in further analytical work. But it is important that the indicators be distinguished from the results of that further analysis: the most common problem with policy monitoring is the tendency of the casual user to read more significance than warranted into the reported level of the indicator. Such policy indicators should be thought of as necessary building blocks for models and evaluations, not as an end in themselves.
The reason to monitor policies is likely to vary by country. For developed countries, much of the interest surrounding policy monitoring has focused on the impacts on trade. For developing countries, the key issues are more likely to be on domestic performance of the agricultural sector. The main questions that such a monitoring system would address would be the extent to which the polices of the country concerned are influencing, in a positive or a negative way, the performance of the sector. Subsequent uses of the quantitative indicators derived in the study could include the trade impacts of policy change or the impacts on producers and consumers in the country of trade policy changes. But the main objective of the monitoring exercise itself is to benchmark the set of policies currently in place.
The definition of monitoring used here implies that the answer to the question “what should one monitor” is itself crucially dependent on the end-use of the indicators. For present purposes we assume that the objective of FAO (and the Consortium) is to provide countries with information that will be useful in their attempts to improve agricultural policies. This suggests that the policies that are the target of the monitoring are clearly identified and that these policies are under the control of the government concerned. Specifically this means that a “policy” is considered as a conscious act of legislation as opposed to a circumstance or economic condition.1
How should one monitor such governmental actions? The method that seems to best to provide the input into decision making is to measure the effect of the policy on the incentives of the actors in the sector concerned and on their capacity to react to those incentives.2 It is convenient to translate those incentives into a quantitative indicator in the price dimension, as a rate of protection, a rate of assistance or a subsidy equivalent.3 The appropriate measure for capacity is more likely to be investment or spending on capacity increases rather than being represented in the per unit (price) dimension. Other policies operate at the macroeconomic level and will be expressed in terms of overall taxes or subsidies to the sector through the operation of the price system. Regulations also reflect policies and their impact must be considered in any full analysis of the incentives to the agricultural sector.
A variety of indicators are available for the purposes of monitoring in a price dimension the incentive effect of agricultural policies. Six of the most usual measures are shown in Table 1. The most straightforward measure is based on the standard analysis of a tariff: the nominal rate of protection (NPR) is given as the increase in the revenue per unit (producer price) in the presence of the policy relative to that which would be obtained in the base (no-policy) state. The adjusted nominal rate of protection (ANR), also found in tariff studies, measures the change in net income per unit, taking into account any taxes and subsidies on purchased inputs, relative to income per unit in the no-policy state. The effective rate of protection (ERP) takes into account the change in value added per unit of output but relates it to the no-policy value added. The added information gained by tracking changes to value added rather than revenue per unit (adjusted or not for input taxes) comes at a cost in terms of more data: all non-factor inputs have to be accounted for regardless as to whether their prices are affected by the policies under study. The choice is therefore governed by available resources and data and by the set of policies whose effects are being measured.
In agricultural policy, as discussed below, an additional set of indicators not so focused on border tariffs has become useful. As countries have moved from price protection (at the border or through subsidies on outputs or inputs) to income protection, it has been necessary to include direct payments and other forms of non-price subsidy. Particularly in the situation of developed countries, it became evident that the analysis of agricultural support would be incomplete without consideration of the influence of government outlays on farmers' returns in the form of capital grants, input subsidies, and various other transfers involving government expenditures. Thus the nominal rate of assistance (NRA) was developed to record the change in income as a result of both price and direct income support as a proportion of income in the no-policy situation. The effective rate of assistance (ERA) measures the same change relative to value added in the no-policy case. Finally, the producer subsidy equivalent (PSE) measures the transfers through price and non-price policies as a proportion of with-policy income.4
If tariffs are the only protective measure used, and there is complete price transmission from border to domestic prices, the NPR is a good approximation of the impact of protection on domestic producer prices. Principal data sources in this case would be the tariff schedules of the countries concerned, and the specific policy information needed would be minimal.5 Where non-tariff barriers to trade such as import quotas, or export subsidies are in operation, then the tariff-equivalent measure of both tariffs and non-tariff barriers is more appropriate. Thus, even when border measures predominate, an important decision has to be made as to whether to rely on explicit border measures (e.g. tariff schedules), or alternatively to estimate the tariff equivalent of tariffs and non-tariff barriers by a direct price comparison approach.6
However, it is important to note that the NPR measure takes no account of the possible effects of protection for major tradable inputs used in a particular product or production process. Resources move between alternative activities not according to the gross revenue from the sector but in accordance with the return to factors employed in that sector. Thus, even when protection is given mainly at the border, a measure of protection that involves value-added, such as ERP, is useful.
The more inclusive measures of protection were developed in large part because the protection given to agriculture at the border is only one part of the impact of policies. The concepts of NRA, ERA and PSE were developed in an attempt to measure in one integral measure the effect of both price and non-price related interventions on farmers' income.
The exchange rate is explicitly involved in the calculation of both NPRs and ERPs in so far as it is used in comparing domestic to border prices. Therefore a misalignment of the exchange rate can significantly affect the measures of the protection given to agricultural goods. It was concern about the possible exchange rate misalignment in developing countries that induced economists in the 1980s to incorporate some measure of this exchange rate phenomenon in the measures of agricultural protection, and to calculate what is often referred to as the “indirect” rate of protection.7 Moreover, the “domestic” measures that aim to include the effects of policies inside the border are also impacted by exchange rate misalignment. So long as any of the goods purchased by the farm sector are “tradable” their price will tend to reflect exchange rate conditions.8 In practice, the calculation of NRA, ERA and PSE does make use of border price comparisons, and so the exchange rate is crucial to the estimation process.
Less easy to categorize are the possible measures of structural policy, factor market conditions, and regulations regarding health and safety. Conditions in the factor markets are important to agriculture, and the provision of infrastructure is an important aspect of government policy. In practice, limited quantitative evidence is likely to be available on the impact of structural policies. However, as discussed below, some indications of the possible effects of these policies on the farm sector should be available. In most developing countries an important role of government is to identify the main bottlenecks in infrastructure, marketing, regulations on health and standards, water markets, and other factors that affect the competitiveness of agriculture and to design appropriate interventions for their removal. However, making such calculations necessitates investigating the whole production and marketing chains to perform such analysis. As a consequence, it is rarely done in cross-country studies. We suggest below that such a systematic analysis of the food chain be a part of the policy monitoring only in those cases where reliable information, or the means to collect it, is available.
The following discussion illustrates the way in which this menu of measures has been employed in agricultural policy monitoring.
1 An example may make this clear. If the agricultural sector in a developing country is inadequately served by roads to connect with urban markets, we do not identify the under-provision of roads as a "policy". In this case the policy is the investment in roads, and would be monitored as such. The indicator may show that this investment is low by various standards, and quantitative analysis may suggest the benefits from increasing such investment. By the same token, if a country chose to have no protection or intervention in a market this would be monitored as "no policy" rather than a policy to allow market forces to determine prices and trade.
2 Of the alternative methods of monitoring, a purely qualitative description of polices has limited use in policy analysis and in cross-country (or even over time) comparison. Indeed, quantitative monitoring takes as its premise that qualitative information about policies is inadequate in itself, though it is often a useful complement to qualitative monitoring.
3 Some policies would be easier to handle in the volume dimension (quotas, acreage restrictions, etc.) but then these policies would need to be converted to a price measure to be aggregated into convenient indicators.
4 The fact that the PSE, in contrast to the other measures, uses the policy-inclusive base as the denominator, is not a significant difference. The transformation to no-policy base can easily be made if this is problematic. It was chosen originally to allow the intuitive interpretation "x percent of income comes from support". Similarly, the ERA is often calculated with the policy-inclusive value-added as the denominator to avoid having to work with a small or often negative non-policy value-added.
5 This is not always straightforward, however, as there is often a difference between bound and applied tariff rates and different rates for imports coming from different sources. These issues will be taken up later.
6 Based on the identification of the point in the marketing channel from which domestic producer prices and corresponding border-price equivalents can be contrasted, such as the price paid at the buying agency for wheat (usually the flour mill), or at the slaughterhouse for pigs, or at the milk processing plant. It is seldom the farm gate price. After the border price has been transformed into domestic currency, all costs relating to transport, storage, loading and marketing, as well as taxes and subsidies to trade in the commodity in question, must be subtracted from/added to the border price up to the point in the marketing channel where domestic and border prices can be compared.
7 Strictly speaking, indirect protection at least in the Krueger, Schiff and Valdés approach is not confined to exchange rate misalignment alone; it attempts to capture also both the influence of industrial protection and of macroeconomic variables (specifically, the "excess" deficit in the current account) on relative prices from an economy-wide perspective.
8 Conceptually, the return to factors producing exclusively non-traded goods and purchasing no traded inputs (as might happen in low-income subsistence agriculture) is not influenced by the exchange rate, but the purchasing power of those factor incomes will still be influenced by the traded goods sector. In practice, it may make sense to concentrate on tradable goods, as the impact of policies in such a case would be much more direct.