Reducing Inflation
Lessons From Albania's Early Success
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

Contributor Notes

Authors’ E-Mail Address: cmcneilly@imf.org; schiesser.doris@snb.ch

Inflation in Albania fell rapidly once comprehensive stabilization policies and market-oriented reforms were launched, in contrast to other transition economies, where price liberalization was generally followed by persistently high inflation. The early reduction of underlying inflation is confirmed by trimmed mean estimates of core inflation, which use a central portion of the commodity-wise inflation distribution. This also demonstrates the usefulness of estimating core inflation for transition economies. The early success in curbing inflation is attributed to the extremely broad scope of initial price liberalization coupled with key supporting measures. It also gives hope for early recovery from Albania’s 1997 economic crisis.

Abstract

Inflation in Albania fell rapidly once comprehensive stabilization policies and market-oriented reforms were launched, in contrast to other transition economies, where price liberalization was generally followed by persistently high inflation. The early reduction of underlying inflation is confirmed by trimmed mean estimates of core inflation, which use a central portion of the commodity-wise inflation distribution. This also demonstrates the usefulness of estimating core inflation for transition economies. The early success in curbing inflation is attributed to the extremely broad scope of initial price liberalization coupled with key supporting measures. It also gives hope for early recovery from Albania’s 1997 economic crisis.

I. Introduction

In Albania the demise of communism began later and turned out to be more disorderly than in other central and eastern European countries. Albania had among the most centralized and repressive regimes in the world from the mid-1940s until late 1990. Long after economic reforms were initiated in other centrally planned economies, virtually all economic activities in Albania remained controlled by the state.

As in other transition economies, inflation in Albania rose quickly during the early stages of transition, fueled by the conversion of suppressed inflation into open inflation and by passive monetary financing of a growing fiscal deficit. But contrary to the experience of most other transition economies—where price liberalization was generally followed by persistently high or at best moderate inflation—inflation in Albania was brought under control rapidly once the government embarked on a stabilization program in the summer of 1992. Annual inflation was reduced from triple to single digits within three years. We find that most of the increases in the price level from mid-1992 through 1995 can be accounted for by price liberalization and increases in administered prices. This is confirmed by trimmed mean estimates of core inflation, which show that core inflation was curbed early on. We argue that Albania’s uncommon success in achieving an early reduction in underlying inflation to low levels is attributable mainly to the extremely broad scope of early price liberalization. This was supported by complementary policies including (1) measures to foster competition and an early supply response; (2) substantial external assistance; and (3) the implementation of restrictive financial policies.

The paper is organized as follows: Section II details the sweeping price liberalization undertaken and the attendant consumer price developments during 1990-95. Section III describes the consequent changes in relative prices and the consumption basket, and compares them to the experience of other transition economies. Section IV highlights the importance of supporting policies that contributed to Albania quickly curbing inflation following price liberalization. Section V confirms that core inflation was low based on trimmed mean calculations, and demonstrates the value of core inflation estimates for transition economies. Section VI summarizes our conclusions and stresses the importance of the early reduction of inflation to the recovery of output.

II. Price Liberalization and Consumer price developments, 1990-1995

Before 1990, only a few elderly Albanians had ever experienced price increases, since all retail and wholesale prices were fixed centrally and had remained unchanged for more than three decades. All consumer goods were strictly rationed. Private production, even of foodstuffs, was prohibited and gray markets were not tolerated. Contrary to other centrally planned economies, black markets were almost unheard of despite widespread shortages. The first change came in July 1990, when the government authorized the sale of fruits and vegetables in parallel markets at freely negotiated prices and state enterprises were granted limited pricing which quickly led to a parallel market for foreign exchange. This marked the beginning of consumer price inflation in Albania, although the Consumer Price Index (CPI) began to measure changes in the price level only in January 1991 (Table 1).2

Table 1.

Annual Price and Exchange Rate Changes, 1990–95

(Percent change)

article image
Source: Institute of Statistics, Albania

Following the introduction of a parallel market for fruits and vegetables, these items virtually disappeared from the state shops by mid-1991. Liberalization of prices for these goods affected 15 percent of the consumption basket.3 The resignation of Albania’s last communist government in June 1991 was accompanied by economic and political chaos, which was reflected in the CPI. Monthly inflation fluctuated around 5 percent in the first ten months of the year (Chart 1 and Appendix I), as central control over the economy collapsed, additional prices were adjusted, and suppressed inflation became open inflation. Official wholesale prices were adjusted by large amounts throughout the second half of 1991, resulting in large increases in budgetary subsidies to consumers. The official exchange rate was devalued for the first time only in August 1991, from 8 to 25 leks/dollar (Table 2), although the parallel exchange rate had depreciated continuously since mid-1990. Passive monetary financing of the ballooning fiscal deficit, along with some further adjustments to controlled consumer prices, initiated a vicious cycle of depreciation—especially in the parallel foreign exchange market—and inflation. Prices of some non-essential industrial consumer goods were freed in November and large adjustments to energy prices were made late in the year. Average annual inflation reached only 36 percent, but by the end of 1991, consumer prices had doubled compared with a year before.

CHART 1.
CHART 1.

ALBANIA: Monthly Inflation and Exchange Rate

December 1990 – December 1995

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

Source: Institute of Statistics and IMF staff estimates.1/ Defined as actual inflation minus the estimated impact of price liberalization and administered price increases.
Table 2.

Exchange Rate, 1990–95

(Lek/U.S. dollar) 1/

article image
Source: Bank of Albania.

Official exchange rate until July 1992; market-determined unified rate thereafter.

In the first half of 1992, monthly price increases were higher than in the previous year, but declined somewhat as the international community mounted a massive effort to provide and distribute food aid (see Section IV.B). Despite a second devaluation of the official exchange rate from 25 to 50 leks/dollar in January 1992, the official rate continued to lag far behind the rate in the parallel market, which by mid-1992 reached more than 130 leks/dollar.

It was only after a new, democratically elected government assumed office in April 1992 that comprehensive economic reform began. The authorities took steps to establish fiscal and monetary control and introduced comprehensive price, trade, and exchange system reforms in mid-1992.4 In July 1992, the government abolished import and export licenses for all but a few export products, removed most restrictions on foreign exchange for current transactions, and introduced a floating market-determined exchange rate. At the same time, the biggest step in price liberalization was taken as the new government liberalized most wholesale and retail prices—with the exception of key essential consumer items—affecting half of the consumption basket (Table 3.5 The government also adopted a policy of liberalizing or raising the remaining controlled prices to cost-recovery levels over a three-year period. In August 1992, administered prices were raised several fold, bringing many of them close to cost recovery levels. In particular, the price of bread—which had been unchanged for more than 40 years—was raised almost five-fold. This price adjustment for bread accounted for almost one quarter of the August CPI increase, and these price reforms are estimated to have accounted for at least two-thirds of the increase (Table 4), although the freeing of the exchange rate also contributed. The immediate effect of these price changes was that the CPI skyrocketed by almost 50 percent in August and jumped again in October. Since there were no indexation mechanisms in force and financial policies were restrictive, relative prices adjusted quickly, the exchange rate stabilized, and monthly inflation declined rapidly. After an initial jump in the unified exchange rate to 117 leks/dollar, it recovered quickly and remained relatively stable at about 100 leks/dollar through year-end. Reflecting the dramatic developments during the year, inflation reached more than 225 percent on both an annual average and end-year basis.

Table 3.

Time Line of Price Reform

article image
Table 4.

Contribution of Price Liberalization and Administered Price Increases to Monthly Inflation, 1992–95 1/

(percent change)

article image
Source: Albanian authorities and IMF staff estimates.

Liberalization or change in administered price as described in Table 3.

In 1993, the largest increases in the CPI accompanied further price liberalization and increases in important controlled prices during the year. In January 1993, more prices were liberalized—including sugar—and selected administered prices were increased which accounted for about two-thirds of monthly inflation.6 During April-June, monthly inflation was zero or slightly negative, reflecting for the first time seasonality related to fruit and vegetable production as well as seasonal appreciation of the exchange rate.7 In July 1993, the price of bread was doubled, accounting for virtually the entire spike in monthly inflation. The new bread price was subsequently made a ceiling price for state bakeries as bread prices came under downward pressure due to the entry of private bakeries, which are estimated to have satisfied as much as 90 percent of bread demand by 1995. The government intended to abolish the price ceiling once there was sufficient competition to prevent the state bakeries from exploiting their previous monopoly position.8 However, the government repeatedly deferred action, with the consequences described below. In the fall, selected energy prices and controlled rents were increased, adding about one-third to monthly inflation in September. Average inflation was still high at 85 percent in 1993, but 12-month inflation declined to 31 percent by year-end (December/December).

In early 1994, price controls were eliminated for some energy products and controlled rent, while non-energy administered prices were raised, accounting for about one-half of January inflation. In April 1994, electricity and kerosene prices were raised substantially toward cost recovery levels, accounting for at least two-thirds of the spike in monthly inflation. During the rest of 1994 monthly inflation was consistently less than 2 percent, and consumer prices even decreased during two summer months, reflecting a rebound in domestic agricultural production and seasonal appreciation of the exchange rate. Annual average inflation eased to 23 percent and 12-month inflation declined to 16 percent by year-end.

Inflation declined sharply further to single digits during 1995. Several administered prices were increased further in January and April 1995, but had a negligible impact on the CPI. Monthly inflation was again less than 2 percent in most months, and prices fell during the summer months. Annual average inflation was just 8 percent and year-end inflation declined to 6 percent. This good performance came at a price, however, as other adjustments toward cost recovery levels, as well as action on the bread price ceiling, were deferred until after the 1996 general elections.

As of end-1995, the weight of controlled prices in the CPI was less than 20 percent. The remaining price controls applied to “essential” goods or items produced under monopolistic conditions. The weight of bread and flour in the CPI is 12.5 percent, so during the roughly two-year period when the bread ceiling price was not binding, the effective weight of controlled prices was reduced to about 7 percent. But given its importance in the consumption basket, the deferral of action on the bread price was critical.

The first indication that the government had postponed liberalization of the bread price too long came when the ceiling price became binding during the winter of 1994/95, during a serious electricity shortage which increased production costs. The government reacted to reduced bread supplies by extending the price ceiling to all bread suppliers and reducing the price of electricity for bakeries. Bread supplies were restored and cost pressures eased with the end of the energy crisis, but by mid-1995 cost pressures re-emerged. The government reacted by reducing other input prices, reducing the tariff rate on wheat and flour, and setting a price ceiling for flour. Nevertheless, by end-1995 the ceiling had again become broadly binding. Some private bakers temporarily defied the ceiling and raised their bread prices by more than 10 percent. There were also widespread reports that bakeries had reduced the size of loaves by up to one sixth to remain within the ceiling. The issue remained contentious but unresolved at year-end. The bread price was finally liberalized in mid-1996 following the elections.

Had the bread price been liberalized in December 1994, we estimate that inflation would have been about 4 percentage points higher in 1995. The experience with the bread price also illustrates the danger of incomplete price liberalization. By not liberalizing the price during the period when the price ceiling was not binding, the government both incurred and imposed substantial costs when it tried to enforce the price ceiling. The government budget lost revenues and eventually incurred new subsidy costs, and the utilities lost revenues. The bakeries’ profit margins were squeezed and then disappeared, leading to widespread defaults on baking equipment loans from the state banks. At the same time, consumers were not really protected from the underlying cost increase since loaf size was reduced.

The other controlled prices remained well below cost recovery levels at end-1995, and in fact were further away than a year earlier given the postponement of most price increases during 1995. Nonetheless, price subsidies had been sharply reduced by price liberalization, and as of end-1995 covered only a few items, with a weight of about 1.5 percent in the CPI. The budgetary cost of these consumer price subsidies was reduced from 3.5 percent of GDP9 in 1992 to 0.6 percent of GDP in 1995.10 During the three major rounds of price liberalization, budgetary transfers to households were used to cushion the impact and help maintain the consensus in favor of economic reform.11 By end-1995, these compensations had been discontinued or consolidated into base wages.

III. Realignment of Relative Domestic Prices

Price and trade liberalization, as well as changes in the tax regime, resulted in significant changes in the relative prices of many goods (Chart 2). As in other transition economies, this included large increases in the relative price of some staples (such as bread, olive oil, dried beans, cabbage and sugar). There have also been large relative declines for many items regarded as “luxury goods” by the previous communist regime (ranging from bananas, oranges, candy and sunglasses, to televisions and washing machines). Prices have also come down in relative and even absolute terms for some foods which were previously very scarce, including milk, eggs, chicken, honey, watermelon and soap. The magnitude of these relative price movements are clear from Chart 3 and Chart 4.

CHART 2.
CHART 2.

Albania: Relative Price Changes, 1990-95

(1990=100)

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

Source: Institute of Statistics, and IMF staff estimates.
CHART 3.
CHART 3.

Albania: Food Price Developments, 1990-95

(1990=100)

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

Source: Institute of Statistics, and IMF staff estimates.
CHART 4.
CHART 4.

Albania: Non-food Goods Price Developments, 1990-95

(1990=100)

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

Sources: Institute of Statistics, and IMF staff estimates.

The relative price of services initially declined and stayed low through July 1993, as the prices of many state-provided services (e.g., electricity, telecommunications) remained controlled. Prices of services then began a rapid catch-up, however, as some services prices were liberalized and others were adjusted toward cost recovery levels (Chart 5). The relative prices of imported food and non-food goods were also held down by the real appreciation of the exchange rate.

CHART 5.
CHART 5.

Albania: Services Price Developments, 1990-95

(1990=100)

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

Sources: Institute of Statistics, and IMF Staff estimates.

In real terms, the lek is estimated to have appreciated against the U.S. dollar by 106 percent from August 1992 to the end of 1995, based on relative CPI movements (Chart 6).12 This strong appreciation undoubtedly contributed to keeping the underlying inflation rate low, since—based on the large trade deficit and observation of domestic markets—imported goods appear to account for a significant share of domestic consumption.

CHART 6.
CHART 6.

ALBANIA: Exchange Rate July 1992 - December 1995

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

Source: Data provided by the authorities, and IMF staff estimates.1/ Index, June 1992=100. Staff estimates based on provisional direction of trade data; an increase indicates an appreciation.2/ Monthly average of commercial bank rates.

A similar picture emerges if we calculate the internal exchange rate based on simplifying assumptions which would allow us to divide the broad categories of the CPI into traded and nontraded goods.13 A few categories can be clearly identified as belonging to one category or the other; for example, medical care, transportation, and communication are classified as nontraded. Where categories are ambiguous, they are assigned on the basis of what is known about the tradability of major items within the category. This approach yields a share of traded goods in the consumption basket of 57 percent, and suggests that the real exchange rate (measured by the relative movements in the price of nontraded versus traded goods) appreciated by almost 40 percent between end-1992 and end-1995 (Chart 6).

By the end of 1995, the relative price of services reached 15 times the 1990 level, a larger increase than for either food or non-food goods, each of which was about 10 times higher. This pattern in the relative price of services mirrors that observed by De Masi and Koen for many countries of the former Soviet Union.14 These shifts in relative prices are also evident in the changing structure of the Albanian consumption basket between 1989 and 1993,15 although it is clear that the share of services—as in other transition economies—remains low by market economy standards (Chart 7).16 While relative food and non-food good prices have largely stabilized, expenditures on services will continue to increase as more services become available, as the prices of government-owned services are raised to cost-recovery levels (e.g., electricity and urban transportation), and as consumers are forced to pay for electricity and telephone usage.17

CHART 7.
CHART 7.

Albania: Comparison of Consumption Baskets

(In percent)

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

Sources: Institute of Statistics (Albania); Goskomstat (Russia); GUS (Poland); National Statistical Service (Greece); INSEE (France); and IMF staff estimates

There undoubtedly remains a large gap between the domestic Albanian and international price levels, although detailed price data are not available to undertake a comparison of the sortwhich has been made for some other transition economies.18 In the case of commodities, for which prices and the trade regime have largely been freed, this is still consistent with the widely observed positive correlation between per capita income and price levels.19 The gap is even larger for services because of the continued existence of some price controls. As pointed out by Richards and Tersman (1995) for the Baltic states, and by De Masi and Koen (1995) for Russia, convergence toward international price levels is likely to be accompanied by further real exchange rate appreciation.

IV. Why Did Albania Succeed? The Importance of Supporting Policies

Inflation in Albania from 1990 to mid-1992 was fueled initially by the conversion of suppressed inflation (acute shortages of staple goods)20 into open inflation, and by passive monetary financing of the fiscal deficit, which ballooned to more than 50 percent of GDP in the first half of 1992. These pressures fed a vicious cycle of depreciation and inflation. As demonstrated above, however, from mid-1992 through 1994, most of the increases in the CPI can be explained by price liberalization and by administrative price adjustments.

Sweeping price liberalization at an early stage, including liberalization of the exchange rate, meant that producers, traders, and consumers responded to appropriate market price signals from early in the transition process.21 It was particularly important that producer prices were liberalized up-front along with the exchange rate, even though this necessitated budgetary subsidies as some consumer prices remained controlled longer. The combination of producers and importers responding to market prices and consumption subsidies quickly being directed to people rather than goods from early on were vital to the rapid response of both production and consumption to market price signals. Comprehensive early liberalization also meant fewer rounds of relative price adjustments in subsequent years, minimizing the sort of price volatility which could have translated into larger increases in the general price level.22 In the absence of indexation, the economy quickly returned to low underlying inflation following these price adjustments (Chart 1).

While these aspects of price liberalization were necessary to the successful early reduction of inflation, they were by no means sufficient. In fact, Albania’s experience stands out: as stressed by De Masi and Koen, “in the large majority of [transition] countries, price liberalization was followed by lasting, high, open inflation …”.23 In addition to the comprehensive scope of early price liberalization, we attribute Albania’s success to support from: (A) measures to foster competition and an early supply response; (B) substantial external assistance; and (C) sustained restrictive financial policies.

A. Measures to Foster Competition and an Early Supply Response

The early privatization of agricultural production led to a quick supply response. Beginning in mid-1991, farmers spontaneously dismantled the agricultural cooperatives and distributed the land and livestock among themselves. By end-1993, 92 percent of agricultural land had been privatized. Largely as a result of this privatization, total agricultural production increased on average by about 15 percent annually in 1992 and 1993.

At the same time, early privatization of retail trade fostered the rapid growth of retail competition as did the liberalization of external trade. During 1991 and early 1992, most retail shops were formally distributed free of charge or for a nominal fee after being taken over by the people working there. Trade liberalization began in 1990, when the state monopoly on trade was eased with the introduction of import and export licensing. In July 1992, Albania removed all quantitative restrictions on imports. The number of product groups subject to export licenses was progressively reduced,24 and in early 1995 the remaining export licensing requirements were abolished.25

Early retail privatization along with external trade liberalization fostered competition and helped to ensure that price liberalization resulted in price competition rather than an opportunity for state enterprises to exploit their previous monopoly power. Moreover, trade liberalization meant that relative domestic prices quickly responded to world price signals. At the same time, the abolition of export licensing encouraged the supply response to price liberalization since prices were no longer kept artificially low to benefit domestic consumers.

B. External Assistance

The international community (led by the European Union and Italy) mobilized a massive aid effort in response to the collapse of the domestic economy and exhaustion of foreign exchange reserves. Official assistance got underway in 1991 and reached a value of almost 50 percent of estimated GDP in 1992 and most than 15 percent in 1993 (Table 5).

Table 5.

Albania: External Assistance, 1991–95

(millions of U.S. dollars, except as noted)

article image
Sources: Data provided by Albanian authorities and donor community; and IMF staff estimates.

The nominal GDP estimates are subject to a high degree of uncertainty. Ratios to GDP should therefore be treated somewhat cautiously.

Because the assistance was largely in the form of food and commodity aid, rather than financial assistance, it expanded supply as well as demand. The availability of this external aid undoubtedly contributed to price stabilization at an earlier stage than would have been possible had the availability of goods depended entirely on the recovery of domestic production in response to price liberalization and privatization. At the same time, foreign aid included staple goods only through 1992, and they were carefully priced on the domestic market to avoid creating disincentives to the recovery of agricultural production.

C. Restrictive Financial Policies

From the second half of 1992, the newly created central bank was able to start controlling money and credit growth, as rapid cuts in the fiscal deficit made it possible to curb central bank credit to the government (Table 6). Tight fiscal policy was essential to monetary control since central bank credit to the government remained the major source of domestic deficit financing even after the introduction of treasury bills in mid-1994. Hard budget constraints were imposed for the first time on state enterprises, and their access to bank financing was also curtailed.26

Table 6.

Albania: Main Macroeconomic Indicators, 1990–95

article image
Sources: Information provided by the Albanian authorities; and IMF staff estimates.

Excluding official transfers.

Including net repayment of arrears.

End-year broad money divided by annual GDP.

As important as fiscal adjustment was the sharp improvement in private savings from 1993 through 1995. This reflected in part positive real interest rates from early 1993, which contributed to a shift from consumption to savings. Higher private savings were also a response to the uncertainties faced by households during the transition, especially high and rising unemployment.

Public sector wage policy (covering civil servants, the military, and state enterprise employees) served as an additional nominal anchor during the crucial early stages of reform in 1992 and 1993. The success of Albania’s disinflation effort is consistent with the view of Sahay and Vegh (1995) that a nominal wage anchor is essential to the success of a money-based stabilization program in a transition economy. They point out that “it is through higher wage bills (either directly through bank credit or indirectly through state subsidies) that most of the money supply makes its way into socialized economies.”27 Therefore, if both prices and wages are liberalized before effective monetary instruments are in place, the economy will essentially be left with no nominal anchor.28

As of mid-1992, the average real wage of civil servants had been reduced to about half of the 1990 level, while real wages of state enterprise employees were still at their 1990 level. The policy was, therefore, to allow gradual recovery of real wages for civil servants—without exacerbating inflationary pressures—while reducing real wages for state enterprise employees in line with sharply lower productivity. This policy was implemented with considerable success through 1993 (Chart 8). Civil service wages were increased by 25 percent in real terms, while state enterprise wages were brought down to the same level. Since employment in state enterprises was much higher than in the civil service, the wage decline resulted in a small further decline in average public sector wages through end-1993 (Table 7).

CHART 8.
CHART 8.

ALBANIA

Real Government Wages, 1991–95

(Index, Dec. 1992 = 100)

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

Table 7.

Albania: Wages and Employment, 1992–95

article image
Source: Institute of Statistics; and IMF staff estimates.

December 1990 = 100.

In 1994, real wages in the civil service increased by another 25 percent, while state enterprise real wages increased by 6 percent. Fortunately this recovery of wages did not impair the progress toward curbing inflation, in part because the wage increases were partially offset by a 15 percent employment reduction in the civil service, and probably in part because the key instruments of monetary control were effectively in place by that time. A further wage increase in the second half of 1995 was not offset by employment cuts and coincided with a more expansionary monetary policy, with negative implications for 1996 inflation performance (see below).

Another key factor in the successful early reduction of inflation following price liberalization was the virtual absence of indexation in Albania. In the absence of indexation, relative price adjustments took place quickly and the economy returned to low underlying inflation within one or two months of major price adjustments. This contrasts sharply with the experience of countries with widespread formal or informal indexation of wages and prices. The adverse impact of such indexation on disinflation efforts has been well described for Slovenia,29 and in Pujol and Griffiths (1996) work showing that the inertial quality of Polish inflation through 1995 largely reflected widespread indexation, especially of wages.30 Such indexation makes the economy less flexible and increases the costs of curbing inflation in several key ways: (i) needed wage differentiation based on productivity is more difficult;31 (ii) public sector wage indexation often gives rise to additional demands for indexation, e.g., for pensions, unemployment benefits, and social assistance paid by the budget. Such automatic increases without regard to the fiscal situation make it more difficult to control the fiscal deficit;32 (iii) widespread public sector indexation can generate pressures for indexation of private sector wages, rents, etc.; and finally, (iv) with sticky prices and wages it takes considerably longer than the one to two months observed in Albania for price shocks (e.g., increases in administered prices, depreciation of the currency) to be absorbed by the economy.

V. Confirmation of Low Underlying Inflation

One could calculate underlying inflation by always excluding the same, most volatile components of the CPI—in the case of Albania, we have argued that during 1992-1995 these were predominantly price shocks in response to price liberalization and increases in administered prices. An additional approach to calculating underlying inflation would be to abstract from seasonal variation, which tends to be important in transition economies because seasonal food prices represent a large share of the consumption basket.33 This is certainly true for Albania, where there has been a robust seasonal pattern in price behavior related to summer harvests (Chart 9).34

CHART 9.
CHART 9.

ALBANIA Seasonally Adjusted Inflation, 1993-95

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

Sources: Institute of Statistics and IMF staff estimates.1/ Seasonal factors derived using X-11 on CPI (1993-95) excluding administered price changes.

In a transition economy, however, relative prices may change well beyond what can be eliminated by selective exclusion of certain items such as food or even administered prices, or exclusion of seasonal variation. This point is reinforced by recent work suggesting that relative price variability associated with price liberalization has a sizable effect at high inflation during initial liberalization, but that other determinants of inflation are more important in the later stages of transition.35

A more general approach is to exclude those items in the CPI which have the biggest or the smallest commodity-wise inflation in each period, i.e., to ignore the tails of the distribution of commodity-wise inflation and concentrate on what happens to most of the prices. This approach is motivated by observations that individual price series (components of the CPI) tend to exhibit substantial skewness.36 Some of these price adjustments might not only be motivated by measurement errors, but also by seasonal factors or supply side shocks which create misleading movements in the aggregate index. A more robust measure of inflation would, therefore, concentrate on the central portion of the commodity-wise inflation distribution where—by definition—relative prices shift the least. The result is a measurement of core inflation?37

One could calculate core inflation based on different large central portions of the distribution. The most extreme approach would be to calculate the inflation of only the median item—thereby ignoring the distribution altogether. An alternative is to calculate the “trimmed mean,” whereby x percent of each tail of the distribution is removed, and the weighted average of the central l-2x percent is calculated. For example, a 15 percent trimmed mean would be the weighted average of the central 85 percent of the distribution.

We divided the Albanian CPI basket into 16 groups. Nine groups were subcategories of food, beverages, and tobacco—with a collective weight of 72 percent in the CPI basket. The remaining seven groups consisted of non-food goods and services (Appendix III). Monthly estimates for 1993-95 of five alternative measures of core inflation are presented in (Table 8).38

Table 8.

Albania: Alternative Measures of “Core” Inflation 1993–1995

(percentage change each month)

article image
Source: IMF staff estimates based on data of the Institute of Statistics.

Seasonal adjustment with XI1.

Average of central 87.5 percent of distribution-using CPI weights-of monthly inflation.

Average of central 75 percent of distribution-using CPI weights-of monthly inflation.

Average of central 50 percent of distribution-using CPI weights-of monthly inflation.

Average of central 25 percent of distribution-using CPI weights-of monthly inflation.

Average of central 12.5 percent of distribution-using CPI weights-of monthly inflation.

Excluding meat, poultry, and fish; dairy products and eggs; oils and fats; fruits and vegetables.

For all five alternatives, average monthly inflation for every year is lower than the average inflation of the entire basket, the seasonally adjusted CPI and inflation excluding selected or all food items. However, which of the five alternatives is the best? The most important selection criteria are: (i) low volatility (core inflation should not bounce around as much as the overall index); and (ii) a high degree of correlation with movements in the overall index (to avoid choosing items whose prices hardly change and to have a good predictor of future inflation). Table 9 shows that the 25 percent trimmed mean (average of the central 75 percent of the monthly inflation distribution) has the lowest volatility relative to its own mean, followed by the 12.5 percent trimmed mean, while the correlation of the 12.5 percent trimmed mean with the overall index is slightly higher than for the 25 percent trimmed mean. The 87.5 percent trimmed mean—in this disaggregation the median—has the lowest absolute standard deviation, but its correlation with the overall basket is only 50 percent. In comparison, the seasonally adjusted CPI is slightly less volatile than the overall index, but its absolute and relative standard deviation are larger than those of the 12.5 and 25 percent trimmed mean. The CPI excluding selected or all food items is much more volatile than the overall index—which is the opposite of what one finds in most market economies, but is consistent with the relative price adjustments taking place in transition economies.39

Table 9.

Albania: Comparison of Alternative Measures of Monthly Inflation

article image
Source: Staff calculations based on official CPI data.

Chart 10 shows monthly core inflation—defined as the 25 percent trimmed mean—against actual inflation. By virtue of its construction, core inflation has lower peaks than actual inflation. Both average monthly core and actual inflation declined after 1993 and began to converge: in 1993, 1.5 percent core monthly inflation versus 2.3 percent actual monthly inflation; in 1994, 0.6 percent versus 1.3 percent; and 1995, 0.4 percent versus 0.5 percent. While in 1993 annual core inflation was 20 percent compared with actual inflation of 85 percent, in 1995 annual core inflation was 6 percent compared with an actual annual average of 7.8 percent. This supports our earlier finding that most of the realignment of relative prices was achieved by 1994.

CHART 10.
CHART 10.

ALBANIA Actual and Core Inflation, 1993-95

Citation: IMF Working Papers 1998, 078; 10.5089/9781451956313.001.A001

It is noteworthy that core inflation showed an upward trend during the final quarter of 1995, rising to about 10 percent (well above recorded inflation of about 6 percent). Thus core inflation captured the effect of both deferral of administered price adjustments, and a relaxation of wage and monetary policy around mid-1995 in the run-up to general elections in 1996, and correctly foreshadowed an upturn in CPI inflation which only became fully apparent around mid-1996 when CPI inflation rose to 15 percent. This experience demonstrates the value of a core inflation index as an indicator of inflation as a monetary phenomenon, abstracting from both real side shocks (including seasonal variation) and the existence of administered prices and measurement errors that may overstate or understate the true change in the overall price level.40

A number of countries produce as a guide for monetary policy estimates of underlying inflation which exclude the most volatile CPI components. For example, the series used by the U.S. excludes food and energy prices; the U.K. excludes mortgage costs; and Canada excludes all three. As noted above, however, in an economy in transition there are many more shifts over time in relative prices than can be captured by excluding any one item (such as food), or even—as we did initially in the case of Albania—by excluding a set of administered prices. At the same time, such fluctuations can give rise to misleading pseudo-turning points.41 These concerns highlight the value to policy makers of a comprehensive measure of core inflation such as a trimmed mean estimate.

The tails of the distribution which are excluded from the calculation of core inflation can also provide useful information to policy makers. If a particular product or group of products consistently shows up in the right tail—that is, with unusually large monthly price increases— which are not explained by administered increases or seasonal factors, it may signal a market imperfection such as the exercise of de facto monopoly power. Such information can help in establishing structural reform priorities.

V. Conclusions and Implications for the Future

As in other transition economies, inflation in Albania during 1990 through mid-1992 was fueled by the conversion of suppressed inflation into open inflation and by passive monetary financing of the ballooning fiscal deficit. But Albania thereafter achieved remarkable success in reining in inflation, from high triple-digit rates in mid-1992 to single digit inflation by mid-1995. This stands in sharp contrast from the experience of many transition economies where initial price liberalization was followed by prolonged periods of high or at best moderate inflation.

From mid-1992 through 1994, most of the increases in the CPI can be explained by price liberalization and by administrative price adjustments. These price changes resulted in dramatic changes in relative prices in Albania, which mirror the general pattern of relative price realignment observed in other transition economies. In the absence of indexation, the economy quickly returned to low underlying inflation following these price adjustments. The early attainment of low underlying inflation is confirmed by trimmed mean estimates of core inflation. Such a core inflation index is a valuable policy tool for a transition economy, since it abstracts from both real side shocks (including seasonal variation) and the existence of administered prices and measurement errors that may overstate or understate the true change in the overall price level. It is thus a far more useful measure for a transition economy than one constructed just by eliminating selected volatile factors (e.g. food, energy) as is done for many industrial countries.

We attribute the rapid reduction of underlying inflation in Albania to the extremely broad scope of early price liberalization, including the exchange rate. This introduced appropriate market price signals from early in the transition process, and necessitated fewer rounds of subsequent relative price adjustments. While sweeping price liberalization was a necessary condition, it was no means sufficient, and we have described how the early success in taming inflation also depended on key supporting policies. (1) Rapid, comprehensive price liberalization coupled with agricultural privatization revived output, while the privatization and liberalization of retail and external trade fostered price competition. (2) Substantial external assistance expanded supply more rapidly than would have been possible based just on the recovery of domestic production. And (3) restrictive financial policies were implemented steadfastly, including sharp reductions in the budget deficit, strict control over money growth, and an effective nominal wage anchor at the crucial early stage of reform.

Reducing inflation quickly was important because it permitted an early, rapid resumption of growth to an average of 9 percent per year during 1993-96. Fischer, Sahay and Vegh (1996) found strong evidence that “growth requires stabilization, and stabilization leads to growth”.42 Albania remains the poorest country in Europe and the devastating economic and social crisis in 1997 further widened the gap. A resumption of sustained, rapid growth will be essential to approach European levels of income. If Sarel (1995) and Ghosh (1997) are correct that the effects of inflation on growth are nonlinear and the critical threshold is around 8-10 percent per annum, then for Albania to achieve the required rates of growth it is vital that the authorities regain the low rate of underlying inflation achieved by 1995. Fortunately, the experience from the early years of transition gives hope that this can be achieved quickly.

Appendix I. Albania: Monthly Actual and Underlying Inflation

article image
Source: Institute of Statistics, and IMF staff estimates.

Appendix II. Albania: Details of the Consumption Basket

article image
Source: Institute of Statistics, Albania.

Appendix III. Albania: Underlying Data for Estimates of Core Inflation, 1992–1995

article image
article image
article image
article image

Appendix IV. Albania: Estimates of −87.5 percent and 75 percent Trimmed Mean Inflation, 1992-1995

article image
article image

Appendix V. Albania: Estimates of 25 percent Trimmed Mean Inflation, 1992-1995

article image
article image
article image
article image

References

  • Brown, Scott, Paul Mathieu, Caryl McNeilly, Reva Krieger, and Caroline Van Rijckeghem, 1994, Albania, IMF Economic Review No. 5 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Bruno, Michael, and William Easterly, 1995, “Inflation Crises and Long-Run Growth,” NBER Working Paper No. 5209 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Bryan, Michael F. and Stephen G. Cecchetti, 1994, “Measuring Core Inflation,” in Monetary Policy, ed. by N. Gregory, Mankiw, National Bureau of Economic Research Studies in Business Cycles, Vol. 29, (Chicago: The University of Chicago Press), pp. 195220.

    • Search Google Scholar
    • Export Citation
  • Coorey, Sharmini, Mauro Mecagni, and Erik Offerdal, 1996, “Disinflation in Transition Economies: The Role of Relative Price Adjustment,” IMF Working Paper 96/138 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • De Broeck, Mark, Paula, De Masi, and Vincent Koen, 1995, “Inflation Dynamics in Kazakhstan,” IMF Working Paper 95/140, (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • De Masi, Paula and Vincent Koen, 1995, “Relative Price Convergence in Russia,” IMF Working Paper 95/54, (Washington: International Monetary Fund)

    • Search Google Scholar
    • Export Citation
  • Fischer, Stanley, Ratna Sahayand Carlos Végh, 1996, “Stabilization and Growth in Transition Economies: The Early Experience,” Journal of Economic Perspectives, Vol. 10 (Spring), pp. 45-66.

    • Search Google Scholar
    • Export Citation
  • Ghosh, Atish R., 1997, “Inflation in Transition Economies: How Much? and Why?,” IMF Working Paper 97/80, (Washington: International Monetary Fund).