Commodity Price Developments Since The 1970s

Presented by: Ali Arslan Gürkan
Chief, Basic Foodstuffs Service
Commodities and Trade Division
Food and Agriculture Organization

Introduction

Since the pioneering work of Prebisch and Singer (1950), which drew international attention to the declining trends in primary commodity prices, economists and policy makers have been intrigued by the causes and consequences of commodity price developments. A rich body of literature already exist on this issue and international actions to cope with such developments have been in place over the last three decades^[13]. The available empirical evidence has produced mixed results regarding what the actual tendencies are: with some showing positive, negative or no decline in commodity price trends. In general, most agree that commodity prices are non-stationary, i.e. they do not revert back to their old levels after receiving a shock, but views differ regarding the nature of nonstationarity, i.e. whether the trend is stochastic or deterministic or whether there are structural breaks - Spraos (1980), Thirwall and Bergevin (1985), Grilli and Yang (1988), Diakasavvas and Scandizzo (1991), Cuddington and Urzúa (1989), Reinhart and Wickham (1994). Nonstationarity in commodity prices has found support recently in empirical evidence provided by Cashing, et. al. (1999), indicating that shocks to primary commodity prices are long lasting with wide variability in persistence levels. Sarris (1998) also found that the underlying trend in cereal prices were deterministic with some tendency of increased volatility during the 1995/96 period. Maizel, Becon and Mavrotas (1997) concluded that commodity prices exhibit a long run decline rather than volatility.

Variability in commodity prices arises as a direct consequence of shocks in underlying demand and supply conditions, and is affected significantly by policy measures implemented at the national level. For storable commodities, Deaton and Laroque (1992) characterize the nature of price shocks depending on prevailing market conditions. In tight markets, a sudden increase to consumption induces a sharp rise in prices which are temporal in nature, while in slack markets, the impact of a shock induces the release of stocks and other policy measures that dampen price increases but render them persistent. They conclude that this behaviour of prices results in price cycles often observed with flat tails and sharp spikes.

During the 1996 FAO meeting of experts on agricultural price instability^[14], it was generally agreed that compared to the past, world commodity markets in the future were likely to be characterized by lower levels of overall stocks, although, at the same time, being less prone to instability because of faster and broad-based adjustments to production/demand shocks. However, the path to the new market environment was seen as uncertain and it was generally felt that price instability would be greater during the transitional period.

The primary aim of this document is to summarise the behaviour of agricultural commodity prices over the last three decades to form the back-drop of discussions in the 2002 Commodity Consultation.

The first part of this note presents a summary view of monthly developments of representative international prices of 18 important agricultural commodities over the period covering 1970-2000. This is done by splitting the period into three decades in an arbitrary manner and presenting the results of statistical analyses conducted to test for differences in levels and variances of the nominal and real prices. The second part of the note reports on the results of more sophisticated analysis of trends and volatility of the same set of prices.

Have there been changes in the levels and variability of international commodity prices over the past three decades?

Appendix I presents a graphical display of both nominal and real prices for all commodities considered in this study. In general, over the last three decades, agricultural commodity prices could be characterized as exhibiting a large shock during the early 70s followed by a series of smaller shocks and relatively flat surfaces.

One simple way of looking at different aspects of price developments is to divide the various series into a number of periods and compare the changes that have occurred in order to discover consistencies across time and across commodities. In this case the series have been divided into three equal samples (arbitrarily), each representing a different decade (i.e. 1970-80, 1980-90 and 1990-2000). Two aspects of the series, both in nominal and real terms^[15], levels and variances, were then subjected to statistical tests^[16].

Table 1 presents the results of the analyses and indicates that the mean real prices of all commodities over the three decades, with the exception of banana during the earlier part of the period under study, appear to have been on the decline.

Table 1: Changes in mean levels (real) over time

All results are significant at the 10% level

Somewhat different results are obtained when nominal prices are analyzed in the same manner. Table 2 indicates that the mean prices of 12 commodities have increased in a statistically significant manner over the first two decades, with 3 experiencing a significant decline and a further three no change. The comparisons over the last two decades, on the other hand, indicate a complete reversal: with the mean prices of 12 commodities decreasing significantly, while 3 increasing and a further 3 exhibiting no change.

Table 2: Changes in mean levels (nominal) over time

(a) - at the 10% level of significance, the mean in decade 2 = mean in decade 3.
(b) - at the 10% level of significance, the mean in decade 1 = mean in decade 2.
(e) - at the 10% level of significance, all means are equal.
All the other results are significant at the 10%.

It is obvious that during the 1990s most of the commodities analyzed have experienced depressed global markets, regardless of the manner in which the prices are measured. More analysis is, however, needed to determine the underlying causes for such a state of affairs. Discussion to be held during the Consultation will hopefully provide guidance for such an undertaking.

Tables 3 and 4 present a summary of the results of the analyses related to changes in variability in real and nominal prices over time as revealed by the estimates of interquartile range^[17] for the three decades. From the latter table, it is apparent that 9 out of 18 commodities (i.e. in nominal terms) showed a lower level of variability in decade 3, relative to decade 2. However, in real terms, 16 out of 18 commodities showed a decrease in variability for the same period. Where there were ambiguities in interpreting changes in variability, a comparative test was conducted using a two sample test for variances at the 5% level. For both real and nominal commodity prices, the following commodities exhibited a decline in variability during decade 3 relative to decade 2: sunflower meal, maize, palm oil, soybeans, soybean meal, rice, cocoa, sugar, and rubber.

Is it possible to look at price variability in other ways?

Another way of looking at the variability of prices is by removing any consistent component of the price series and analyzing the resulting variability of the residuals. The analysis was carried out by estimating a stochastic trend unobserved component model (for each of the 18 commodities in both real and nominal terms) using the State Space Kalman filter procedure (see Appendix VI for details) and then testing to see whether the resulting distribution of residuals is significantly different from a normal one. In this context, the departure from normality is viewed as being a measurement of volatility caused by factors that could not be represented by the consistent components of the processes generating the time series. The tests tend to support the analyses conducted using more traditional measures of variability/volatility, in that, in real terms, 11 out of 18 commodities (cocoa, maize, palm oil, rapeseed, rubber, soybean, soy meal, sugar, sunflower, rubber, and tea) show less volatility over the last two decades of the period under study: Though not to the same extent^[18].

Table 3: Changes in real price variability over time

Other tests were also conducted to test for the stability of a different kind: the stability of the reduced form representation of a number of commodities where time series data are available for all the main variables relevant for the markets, i.e. consumption, production and ending stocks. This was done by testing for the stability of the coefficients of a reduced form price equation. The coefficients were estimated recursively and their behaviour examined. For the subset of commodities, it was observed that the structural parameters were stable throughout the last decade of 1990s, indicating no significant behavioural changes affecting the response coefficients.

Is there empirical evidence to suggest structural breaks in the price series over the past three decades?

Very little attention has been given to the issue of structural changes in agricultural commodity prices. As the agricultural sector is intertwined with other sectors and constitutes a major contribution to economic activity in many importing and exporting countries, economy-wide changes in the levels of economic activity would have a direct impact on the agricultural sector. Besides, changes in the economic structure, agriculture is perhaps more prone to shocks caused by weather and other natural disasters, which can have sustained and lasting effects. In addition, technological changes can alter productivity levels and can shift the way resources are allocated, thus, leaving permanent effects on the agricultural sector. In addition, major policy reforms both at the national and international levels can induce structural changes in prices.

Table 4: Changes in nominal price variability over time

Using the full series available, and assuming that break points are not known a priori, the outlier procedure was used (see Yin and Maddala, 1997 for full details) to identify the year or years in which breaks could have occurred. The rationale of the procedure is that outliers are aberrant observations that are away from the rest of the data, caused by errors in measurements or unusual events such as changes in economic policies, wars, disasters, and so on (Perron, 1989). Structural change outliers are classified as (i) additive outliers (one-time changes) - these occur when there is a spurious change in the data, but the data returns back to its normal pattern in subsequent periods; (ii) level shifts - occur when the effect of a large innovation persists over time. As can be seen from Table 5, for the important basic food commodities 1988 appears to be a year in which persistent level changes have occurred, though not all seem to have had a significant break during the food crisis of early 1970s^[19]. For tea and cocoa, early 1980's appears to be a period when level breaks occurred.

Table 5: Probable structural changes for selected (real) commodity prices

Break dates were significant at the 1% level.

Additive changes refer to one time break and level changes refer to changes that persist over time.

References

Cashing, P. and Pattilo, C (2000). "Terms of Trade Shocks in Africa: Are they Short-Lived or Long-Lived?", IMF Working Paper, WP/00/72.

Cashing, P., C. J. McDermott and Alasdair Scott, (1999). Booms and Slumps in World Commodity Prices", IMF Working Paper, WP/99/155

Cashing, P., Hong Liang and C. J. McDermott, (1999) "How Persistent Are Shocks to World Commodity Prices?", IMF Working Paper, WP/99/80.

Cashing, P., C. J. McDermott and Alasdair Scott, (1999). "The Myth of Comoving Commodity Prices", IMF Working Paper, WP/99/169.

Cashing, P., C. J. McDermott, (2001). "The Long-Run Behavior of Commodity Prices: Small Trends and Big Variability", IMF Working Paper, WP/01/68.

Cuddington, J. (1992). "Long-Run Trends in 26 Primary Commodity Prices: A disaggregated look at the Prebisch-Singer Hypothesis", Journal of Development Economics, vol. 39, pp.207-227.

Cuddington, J. and H.C. Wei (1990). "Trends and Cycles in the Net Barter Terms of Trade: A New Approach", Economic Journal, vol.99, pp. 426-442.

Deaton, A. and Guy Laroque, (1992). "On the Behaviour of Commodity Prices", The Review of Economic Studies, vol. 59, pp. 1-24.

Engle, R. (1995). ARCH: Selected Readings. Oxford University.

Grilli, E. and M.C. Yang (1988). "Primary Commodity Prices, Manufactured Goods Prices and the Terms of Trade of Developing Countries: What the Long Run Shows", The World Bank Economic Review, vol. 2, pp.1-47.

Hamilton, J. D. (1994). Time Series Analysis. Princeton University Press, NJ.

Harvey, A. (1989). Forecasting Structural Time Series Models and the Kalman Filter. Cambridge University Press.

Harvey, A. (2000). "Trend Analysis." Faculty of Economics and Politics, University of Cambridge.

León, J. and R. Soto (1995). "Structural Breaks and Long-Run Trends in Commodity Prices", The World Bank Policy Research Working Paper, WPS 1406.

Madala, G. S. and Kim, I. (1998). Unit roots, Cointegration, and Structural Change. Cambridge University Press.

Maizels A., Bacon, R. and George (1997). Commodity Supplie Management by Producteur Countries: A Case-Study of the Tropical Beverage Crops (Unu/Wider Studies in Development Economics)

Powell, A. (1991). "Commodity and Developing Country Terms of Trade: What Does the Long Run Show?", Economic Journal, vol. 101, pp.1485-1496.

Prebisch, R. (1950). "The Economic Development of Latin America and its Principal Problems". Economic Development for Latin America United Nations, NY, USA.

Reinhart, C. and P. Wickham (1994). "Commodity Prices: Cyclical Weakness or Secular Decline?", IMF Staff Papers, vol. 41, No. 2., pp. 175-213.

Sarris, A. H. (2000). "Has World Cereal Market Instability Increased?". Food Policy, vol. 25, pp. 337-350.

Sarris, A. H. (1998). "The Evolving Nature of International Price Instability in Cereals Markets", Food and Agriculture Organization of the United Nations, Commodities and Trade Division, Rome.

Sarris, A.H. (2000) "World Cereal Price Instability and a Market Based Instrument for LDC Food Import Risk Management", Food Policy, vol. 25, pp. 189-209.

Sapsford, D. (1985). "The Statistical Debate of the Net Barter Terms of Trade between Primary Commodities and Manufactures". The Economic Journal, vol. 95, pp.781-788.

Spraos, J. (1980). "The Statistical Debate on the Net Barter Terms of Trade between Primary Products and Manufactures". The Economic Journal, vol. 90, pp. 109-128.

Stock, J. and Watson, M. (1988), "Testing for Common Trends", Journal of the American Statistical Association, vol. 83, pp. 1097-1107.

Food and Agricultural Organization. (1996). Report of a Meeting of Experts on Agricultural Price Instability. Commodities and Trade Division.