Productivity, net returns, and efficiency land and market reform in vietnamese rice production

Productivity, Net Returns and Efficiency: Land and Market Reform in Vietnamese Rice Production By Tom Kompas Crawford School of Economics and Government Australian National Universit

P

09-02

International and Development Economics

H AUSTRALIAN NATIONAL UNIVERSITY

Crawford School of Economics and Government

WORKING PAPERS

roductivity, Net Returns and Efficiency:

Land and Market Reform in Vietnamese Rice Production

Tom Kompas Tuong Nhu Che

Ha Quang Nguyen Hoa Thi Minh Nguyen

Crawford School of Economics and Government THE AUSTRALIAN NATIONAL UNIVERSITY http://www.crawford.anu.edu.au

Productivity, Net Returns and Efficiency:

Land and Market Reform in Vietnamese Rice

Production

By

Tom Kompas Crawford School of Economics and Government

Australian National University

Tuong Nhu Che Australian Bureau of Agricultural and Resource Economics

Ha Quang Nguyen Forestry University of Vietnam

Hoa Thi Minh Nguyen Centre for Analysis and Forecasting Vietnam Academy of Social Sciences

ABSTRACT

Extensive land and market reform in Vietnam has resulted in dramatic increases in rice output over the past thirty years The land and market reforms in agriculture were pervasive, moving the system of rice production from commune-based public ownership and control to one with effective private property rights over land and farm assets, competitive domestic markets and individual decision making over a wide range of agricultural activities The effect of this reform period and beyond is detailed with measures of total factor productivity (TFP), terms of trade and net returns in rice production in Vietnam from 1985 to 2006 Results show that TFP rises considerably

in the major rice growing areas (the Mekong and Red River Delta areas) during the early years of reform, and beyond, but also that there is clear evidence of a productivity “slow-down” since 2000 The differences over time and by region speak directly to existing land use regulations and practices, suggesting calls for further land and market reform To illustrate this, additional frontier and efficiency model estimates detail the effects of remaining institutional and policy constraints, including existing restrictions on land consolidation and conversion and poorly developed markets for land and capital Estimates show that larger and less land-fragmented farms, farms in the major rice growing areas, and those farms that are better irrigated, have a greater proportion of capital per unit of cultivated land, a clear property right

…fty per cent of the increase in total factor productivity (TFP) during thepeak of the reform period (Che et al 2006) Overall, given these reforms,Vietnam has gone from being a large importer of rice in 1976-80, to now thesecond largest exporter of rice in the world, with considerable increases infarm pro…tability and rural incomes and resulting rural poverty rates falling

by over sixty percent from 1993 to 2004 alone (Hansen and Nguyen 2007).However, much still remains to be done to increase living standards

in rural areas and enhance general rural development Like many reformprocesses, the early rapid increases in economic activity have dissipated overtime, with the suggestion now of a TFP ‘slow down’ in rice production inmany areas in Vietnam In addition, many of the poor still farm small areas

of land, constrained in use, often with fragmented or non-contiguous plotsand with little or no human and physical capital accumulation or access toagricultural extension services Much of this is due to remnants from pastinstitutional arrangements, but also to continued constraints in land use,credit availability and the provision of rural services, all calling for further

or renewed land policy and market reform

This paper has two basic tasks First, it assembles a data set from 1985

to 2006 to measure the changes in TFP, terms of trade and net returns inVietnamese rice production, both in the principal rice growing areas andthroughout the country The results track the e¤ects of the major landand market reform process, and beyond, and determine key di¤erences inTFP and net returns over time and by region All of this speaks directly

to existing land use practices and is suggestive of needed policy response.With this is mind, the second task is to isolate the remaining institutionalconstraints and policy challenges that may be limiting increases in produc-tivity and e¢ciency For this purpose, three di¤erent stochastic frontier andine¢ciency models, with varying samples and levels of aggregation, are esti-mated to determine the potential e¤ects of ongoing issues over land use andthe provision of credit, land fragmentation, less than secure property rightsand the lack of rural education and support services

Section 2 of the paper provides context, highlighting the nature and

extent of the past market and land reform process and the remaining tutional barriers and policy challenges in land use practice and rural develop-ment Section 3 brie‡y summarizes the various data sets used in the paper,along with variable de…nitions and econometric speci…cation An extensivedata appendix details the sources of the data, as well as data constructionsand various adjustments Section 4 provides the measures of TFP, terms oftrade and net returns in rice production, while section 5 provides estimatesfor the stochastic frontier and ine¢ciency models Three di¤erent modelsare estimated, in part due to limitations in the data, and in part to highlightdi¤erent aspects of the constraints on ine¢ciency Here, estimates for theprimary provincial data set (1991-99), a period over which there is relativestability in estimated input coe¢cients, is augmented with estimates drawnfrom two farm and household surveys for the year 2004 Section 6 concludes.

insti-2 Context

Rice continues to dominate agricultural production in Vietnam, with nearly

73 per cent of the population still living in rural areas, and rice accountingfor nearly 90 per cent of the output of food grains and almost two-thirds

of rural farm income (SDAFF 2006) Although rice is produced in everyone of the 60 (recently de…ned as 64) provinces in Vietnam, the Red RiverDelta (RRD) and the Mekong River Delta (MRD) are the main rice growingregions The smallest producers of rice (less than 100,000 tons per year) are

in Binh Phuoc province, which is relatively small in area, and contains theprincipal co¤ee growing (Gialai Kontum) and mining (Cao Bang, Bac Kan)areas Provinces with the largest rice output (more than a million tones peryear) are located in the MRD (Tien Giang, Soc Trang, Long An, Kien Giangand An Giang), which as a whole accounts for roughly half of Vietnam’soutput of rice and most of its international exports In terms of naturalconditions, the MRD and the RRD are the most favorable for growing rice,with many areas naturally irrigated, producing up to three rice crops peryear Based on farm survey data for 2004 (as used below), the average farmsize in the RRD (0.4 hectares per farm) is much smaller than in the MRD(1.4 hectares per farm) However, the number of threshing machines in theRRD is almost double that of the MRD In the MRD, with a large volume

of high-quality rice exports, rice processing takes place in mills rather than

on the farm to maintain international standards

As mentioned, rice output has increased dramatically during the majormarket and land reform periods, and by more than three times nationwide(10 million to nearly 34 million tons) from 1980 to 2004 (Kompas 2004) Af-ter a period of ‘output share contracts’ (from 1981-87), as a highly tentativemove to limited property rights and domestic markets, a period of ‘tradeliberalization’ (1988-94) brought major institutional change, allowing for ef-

fective private property rights over both land (initially 10–15 and later 20year leases) and capital equipment With reform, most production decisionswere de-centralized, all farm income (after tax) was retained by the farmerand rice could be sold freely in competitive domestic markets The resultwas an increase in rice prices (from the state controlled low prices prior

to reform) and added pro…tability, which not only increased TFP ably (Che et al 2006) but also generated dynamic gains from trade reformfrom induced capital accumulation out of retained farm earnings (Che et al.2001)

consider-Since 1994, these dramatic market and land reforms have been solidi…edand in many cases extended in the ‘post-reform’ period (For a detailed dis-cussion of land policy reform in Vietnam, see Chu (1992), Fforde (1996), andMarsh and MacAulay (2002).) Nevertheless, a number of concerns remain,and have been raised again recently in a ‘Participatory Poverty Assessment’(PPA), with over 240 focus groups and 1,450 participants, undertaken bythe Vietnam Academy of Social Sciences (VASS 2009) Chief among theseare issues surrounding land title and use, land fragmentation and the lack

of rural credit availability and supporting rural services, and especially so(for this paper) as they impact on productivity and e¢ciency gains

Land title and use requirements are a good example of the challenges thatremain Although the Land Law of 1993 (amended and clari…ed in 1998,

2001 and 2003) allows “land use rights to be transferred, exchanged, leased,mortgaged and inherited” (Congress of Vietnam 1993), in practice consid-erable constraints remain in place regarding both land conversion (i.e., landtransferred or converted to other uses) and land accumulation (i.e., tradesand accumulation of land plots) For any land conversion, for example, thecommune authorities have to …rst develop a plan, often based on or as mi-nor amendments to past historical ‘blueprints’ for land use in that area, tosubmit to various levels of government for approval The PPA reports thatthis process is often long and that transactions costs are high, making itdi¢cult for poor farmers in particular to participate (VASS 2009) In ad-dition, although land consolidation in Vietnam is occurring, with a number

of important bene…ts (see Ravallion and van de Walle 2008), there are stillrestrictions on overall land size In 2007, the Vietnamese government in-creased limits on the transfer of land use rights for annual agricultural landfrom 3 to 6 hectares for the South East, Mekong River Delta and Ho ChiMinh City and from 2 to 4 hectares for other cities and provinces This

is a welcome albeit modest change for many farmers, but in most casesrice farming outside of the MRD is still takes place on very small farms, atsubsistence levels (GSO (VHLSS) 2004, VASS 2009)

Part of the obstacle to land consolidation is the lack of fully secureproperty rights Land use title for agricultural land was extended from

15 to 20 years with the Land Law of 1993, but in many cases even 20years is too short to provide secure rights in the shift to larger farms, or

to a process where farm land is turned into use for small manufacturing

or industry Overall the process of land certi…cation or entitlement itselfhas also been below expectations Household survey data from 2004 (GSO(VHLSS) 2004) suggests that only 76 percent of agricultural land parcels,

68 percent of urban land parcels and 34 percent of forest land parcels havebeen granted land-use right certi…cates In practice, this means about twothird of the total parcels of Vietnam still lack a certi…cate (World Bank2009) Even where land certi…cates do exist, there is a shortage of basicinfrastructure for an e¤ective operation of land administration, includingcadastral mapping, transaction registrations and record management in theprovision of land administration services Problems are compounded bylack of information or public awareness and the apparent limited capacity

of land administration sta¤, especially at the commune and district levels(World Bank 2009) It is perhaps for this reason that real estate marketshave been slow to develop, with recent data indicating that the role of landrental markets in agriculture in rural areas remains thin (GSO (VHLSS)2004), and that continued government restrictions often prevent low costand e¢ciency enhancing transfers (Deininger and Jin 2008) Land rightsthat are not secure also clearly impact on credit availability and capitalaccumulation The PPA reports that farms without land-use certi…cates,which would normally be used as collateral, are not able to obtain even shortterm loans, much less transfer land use entitlements This is also often truefor farms with land tenures that are close to expiry, as most currently are,given the 20 year leases initiated in 1993 (VASS 2009)

Land fragmentation occurs when farms have land use rights to a ber of often small, non-contiguous plots With reform and the resultingdissolution of the commune system in Vietnam, land was allocated to priorcommune members in a roughly equalitarian manner: equal numbers of plotsper household, with a distribution of plots throughout the commune so that

num-no one household would have a concentration of plots in the most fertileparts of the commune, or near roads, water sources or other essential ser-vices Immediately after the major reform process (1988-94), there were asmany as 75 to 100 million parcels or plots of land in Vietnam and on averageabout seven to eight plots per household (World Bank 2003 and Hung et

al 2007), of which about 10 per cent of these plots had an area of only 100square meters or less (Phien 2001) Evidence suggests that plot numbershave been falling recently with land consolidation (nationwide, falling to 4.3plots per household with the fall in the north from 6.0 to 4.9 plots (GSO(VHLSS) 2004), but the problem is still common Fragmentation levels, forexample, continue to remain high in the RRD, the most populated region,with 90 percent of the households having rice farm sizes of roughly 0.2 to0.5 of a hectare (Chu 2008) and an average of 4.6 plots per farm (GSO(VHLSS) 2004) The number of plots per farm in the MRD, by contrast,

is only 1.6 In some cases (e.g., risk spreading), fragmentation may be an

advantage (see Marsh et al 2006), but for the most part small and scatteredland holdings hamper mechanization, the use of bu¤aloes and tractors andtechnological adaptation, as well as generate additional time and labour forfarming activities that must have been carried out in geographically distantplots The embankments that separate plots alone have been estimated toreduce total agricultural land for cultivation in Vietnam by 2.4 to 4 per cent(Thanh 2008).

3 Data sources, variables and speci…cations

Four di¤erent data sets are used in this paper The …rst is a provincial leveldata set on rice production in Vietnam, from 1985 to 2006, for 60 provinces,used to construct TFP, price and quantity indexes and net return measures,greatly improving on the basic TFP estimates (to 1994) provided in Che

et al (2006) This is an extensive data set on prices and quantities forall rice outputs and inputs, directly obtained as or aggregated to provincialaverages The key variables include paddy rice output, labour, land, materialinputs (especially fertilizer, but also seeds and pesticide), capital (a measure

of tractors and bu¤aloes), as well as input prices for labour, land, fertilizer,pesticides and capital used in rice production (See the data appendix fordetailed data sources, constructions and adjustments.) The second data set,used to construct a stochastic production frontier and ine¢ciency model, is

a subset of the …rst, or a balanced panel data set for 60 provinces from 1991

to 1999 The key variables are rice output, capital, labour, land, fertilizer(and other material inputs) and measures of farm characteristics by location,size and the proportion of tractors used per area of cultivated rice land Thechoice of using the sub-sample for 1991-99, rather than the entire sample1985-2006, was made based on a series of Chow tests for structural breaks

to test whether input elasticities were constant over time, with Quandt tests con…rming the hypothesis of di¤erent sub-sample variances,especially after the year 2000 The years 1991-99 thus provided the mostconsistent provincial-level data set with relatively stable input elasticities.The third data set, also used to construct a stochastic production fron-tier and ine¢ciency model, is a farm survey data set, for 388 rice farms,conducted in 2004 in major rice growing regions (the MRD and RRD), de-signed especially to isolate the potential e¤ects on ine¢ciency from landfragmentation Key additional variables include measures of soil qualityand irrigation, and the number of plots, as a proxy for land fragmentation,along with level of education of the head of the farm The fourth dataset uses the 2004 Vietnam Household Living Standards Survey (VHLSS) topartly con…rm the results of the smaller farm survey data set, along withproviding added estimates of the e¤ects of secure property rights and access

Goldfeld-to agricultural extension services The VHLSS is a household survey data

set of roughly 9,000 households in 2,216 communes, with cluster-samplingtechniques to cover the entire country, conducted by the General StatisticalO¢ce (GSO) in Vietnam in selected years (i.e., 2002, 2004 and 2006) The

2004 data set, in particular, has separate components for land use and cultural production Sample size is reduced to 3865 households to isolatefarms that are primarily rice producers

agri-For all stochastic production frontiers log-likelihood speci…cations testswere used to determine functional form and the presence of ine¢ciency ef-fects In all cases, standard OLS estimates are shown to be inappropriateand the functional form rejects a translog speci…cation in favor of a morestandard Cobb-Douglas production function Tests on the cross-sectionaldata sets (the farm survey data set and the VHLSS data) also indicate thatestimates of the stochastic frontiers using a random coe¢cients approach,allowing for ‘non-neutral’ shifts in the production frontier, following Kalira-jan and Obwana (1994), resulted in little di¤erence in estimated coe¢cients,with the ine¢ciency term adequately represented by a truncated half-normaldistribution Frontier and ine¢ciency estimates are obtained using Frontier4.1 (see Coelli et al.1998) for the provincial and VHLSS data and Stata 10for the farm survey data, so that distributional assumptions on the technicaline¢ciency terms could be more easily examined

4 Total factor productivity, terms of trade and net returns

TFP is a measure of outputs to inputs over any two time periods Resultsfor Vietnamese rice production are generated using Tornqvist quantity (andprice) indexes given by

for input weighted shares ifor periods s and t For convenience, results aresummarized across eight regions, as o¢cially de…ned in Vietnam: the RedRiver Delta (RRD) (1), the Northeast (2), the Northwest (3); the NorthCentral Coast (4), the South Central Coast (5), the Central Highlands (6),the Southeast (7) and the Mekong River Delta (MRD) (8) As mentioned,the RRD and the MRD are the major rice growing regions in the country.Region 7 is largely industrial, and regions 2, 3 and 6 are the poorest byconventional measures.

There is little doubt that the increase in rice production in Vietnam hasbeen substantial, especially after the ‘output share contracts’ period (1981-87), or under the major land and market reforms (1981-94 and forward).For the country as a whole, the indexed value of paddy rice output shows

an average annual increase of 3.5 per cent, as a …tted linear trend (see …gure1) The largest increases in rice output occur during the period of ‘tradeliberalization’ (1987-94), and continue (virtually unabated) in the ‘post-reform’ period to 2006 However, trends in TFP vary markedly betweenregions in the country Figure 2 shows that not only is TFP higher in theMRD, and especially so compared to regions (2) to (7), but also that thegrowth in TFP is substantially larger in the MRD compared to the RRDand other regions As a …tted annual trend the growth in TFP in the MRD

is 4.42 per cent, while in the RRD it is 2.25 and in all other regions 1.36 percent per year This poor TFP performance is added concern in the poorestregions of the country (regions 2, 3 and 6), where the average annual increase

in TFP is less than 1.3 per cent In total, the MRD remains a stand-out inboth the level and growth of TFP

In all cases, except for the MRD, there is also evidence of a ‘slow down’

in productivity after the year 2000 This is an added problem, again, in poorregions (which generally do not have a natural advantage in rice production,

or su¢cient water resources for ‘wet land’ rice), but it is also a concern inthe RRD, a major rice growing area, where farms remain relatively smalland fragmented

Figures 3 and 4 build the terms of trade (TOT) in rice production Withthe ‘trade liberalization’ reform process, the indexed price of rice increases(from the state-controlled low price in the ‘communal period’ to the partiallycontrolled price during the output share contracts period) throughout 1989

to 1994 and beyond (see …gure 3) During the trade liberalization periodall controls over the domestic price were removed, and prices rose rapidly.After 1996 the domestic price of rice partially re‡ects world prices for rice

on international markets, and ‡uctuates accordingly

As a product of both rapid increases in the output of rice and economicdevelopment, …gure 4 shows the (roughly uniform) increase in the indexedvalue of input prices over the period Much of this is dominated by increases

in the price of fertilizer (albeit with considerable volatility), due to the rapidincrease in rice output, but farm wage rates also increase at an average

annual rate of 1.44 per cent The TOT, …nally, is summarized in …gure 5, andclearly shows that except for the period from 1989 to 1998 (excluding 1993),the TOT has worsened relative to the 1985 starting point If nothing else,this highlights the importance of TFP increases to partly o¤set this trend,since increases in productivity will generate proportionally more revenuesfor given input use.

Figure 6 is the key graphic, in e¤ect combining all price, quantity andproductivity indexes together It shows both the indexed value of paddyrice output (i.e., the indexed price multiplied by the indexed quantities ofrice) and the indexed value of input expenditures (i.e., the indexed inputprices multiplied by the indexed quantities of inputs) The gap or wedgebetween the two lines provides a measure of ‘net income’ in rice productionover time For Vietnam as a whole, land and market reform generates

a substantial wedge, or considerable increases in net income from 1989 to

2000, and especially so in the years 1992 to 1999 In 2001, both the domesticand international price of rice fall dramatically, and the wedge closes.The wedge for the years 1992-99 provides an essential story of economicdevelopment, and also coincides with well documented falls in the ruralpoverty rate in Vietnam Increases in the price of price, output and produc-tivity, on the one hand, and increases in the farm wage rate on the other,result in substantial increases in farm and rural income It is roughly duringthis period, or from 1993 to 2004, as mentioned above, that the de…ned share

of poor people in Vietnam “dropped by two thirds and approximately 30million people were lifted out of poverty” (Hansen and Nguyen, 2007) Notall of this poverty reduction was due to rice production, of course, but giventhe large share of the population in rural areas and the predominance ofrice production in rural agriculture, there is little doubt that the reform andpost-reform periods had a major impact on overall living standards (SeeGlewwe et al (2002) for the distributional e¤ects of poverty reduction inVietnam, based on early household surveys in 1993 and 1998, and Ravallionand van de Walle (2008) for recent welfare impacts of land reform)

Nevertheless, it is also clear that these gains are not shared equally acrossregions or the country Figure 7 shows the indexed ratio of the value ofrevenues to the value of input costs for paddy rice production, for selectedregions, as a measure of ‘net returns’ Relative to the starting point, allregions due well from 1992 to 1999, but after 1999 both the RRD and allother regions fall (in some cases) far below the starting point For the years1999-2004, net returns are even less than one for areas outside of the MRDand RRD Only the MRD does consistently well, both in terms of levels(compared to all other regions) and in terms of years with relatively largenet returns

5 Frontiers and e¢ciency

The importance of TFP levels and increases in resulting measures of netincome highlight the importance of potential e¢ciency gains that accom-pany further land and market reform The following sections use stochasticproduction frontiers and ine¢ciency models to isolate the key constraints one¢ciency gains (as a component of TFP), and what policy measures might

be most suitable

Stochastic production frontiers were …rst developed by Aigner, Lovell andSchmidt (1977) and Meeusen and van den Broeck (1977) The speci…cationallows for a non-negative random component in the error term to generate ameasure of technical ine¢ciency, or the ratio of actual to expected maximumoutput, given inputs and the existing technology The idea can be readilyapplied to panel data, following Battese and Coelli (1995) Indexing …rms

by i = 1; 2; :::; n; the stochastic output frontier is given by

Yit= f (Xit; )evit uit

(5.1)for time t = 1; 2; :::; T , Yitoutput, Xita (1 k) vector of inputs and a (k 1)vector of parameters to be estimated Cross-sectional estimates (as with thefarm survey data below) drop the index for time, of course As usual, theerror term vit is assumed to be independently and identically distributed

as N (0; 2

v) and captures random variation in output due to factors beyondthe control of …rms The error term uit captures …rm-speci…c technicaline¢ciency in production, speci…ed by

for zit a (1 m) vector of explanatory variables, a (m 1) vector ofunknown coe¢cients and wit a random variable such that uit is obtained by

a non-negative truncation of N (zit ; 2

u) Input variables may be included

in both equations (5.1) and (5.2) as long as technical ine¢ciency e¤ects arestochastic (see Battese and Coelli 1995)

The condition that uit = 0 in equation (5.1) guarantees that all vations lie on or beneath the stochastic production frontier A trend canalso be included in equations (5.1) and (5.2) to capture time-variant e¤ects.Following Battese and Corra (1977) and Battese and Coelli (1993), varianceterms are parameterized by replacing 2

and clearly must have a value between zero and one The measure of cal e¢ciency is thus based on the conditional expectation given by equation(5.3), given the values of vit uitevaluated at the maximum likelihood esti-mates of the parameters in the model, where the expected maximum value

techni-of Yit is conditional on uit = 0 (see Battese and Coelli, 1988) E¢ciency can

be calculated for each individual …rm per year by

E[exp(ui) j vi+ ui] = 1 ( a+ (vi+ ui)= a

1 ( (vi+ ui)= a) exp (vi+ ui) +

2

a=2(5.4)

and ( ) the density function of a standard normalrandom variable (Battese and Coelli 1988) The value of = 0 when thereare no deviations in output due to ine¢ciency and = 1 implies that nodeviations in output result from random e¤ects, or variance in v

As mentioned, the …rst frontier estimate uses a provincial level data set from1990-99 Summary statistics are reported in table 1 Note that all outputand input measures (e.g., average farm size, labour, material inputs) aremultiples of the number of crops per year Generalized likelihood ratio testsare used to help con…rm the functional form and speci…cation As a pre-test, the null hypothesis of a Cobb-Douglas form of the production functionwas tested against a general translog speci…cation by setting the relevantparameters for squared and interaction terms in the translog form equal tozero The resulting test statistic was 2

10= 9:4 compared to a critical value

of 19.7 A Cobb-Douglas functional form was thus selected Accordingly,equation (5.1) for unbalanced panel data set (1991–1999) for i province and

t time period is speci…ed by a production function in log-linear form, or

ln Yit= 0+ 1ln Kit+ 2ln LABit+ 3ln LANit+ 4ln IN (5.5)

+ 5T + vit uit

where Y is the output of rice, K the stock of capital (a combined tractor andbu¤alo measure, in horsepower), LAB labour in working days, LAN totalland under cultivation, IN material inputs (fertilizer, seed, insecticide) and

T is a time trend

The provincial ‘farm-speci…c’ factors used in the technical ine¢ciencymodel, or equation (5.6) below, are average farm size (SIZE), the percent-age of rice land in which tractors are used (T L); a variable indicating soilconditions (SOIL) as a binary variable for the main rice growing regions, orthe MRD and the RRD, the number of threshing machines (M A) and thenumber of tractors (CA), so that

uit= 0+ 1ln SIZEit+ 2ln T Lit+ 3SOILit+ 4ln M Ait (5.6)

+ 5ln CAit+ !it

for !it N (0; !): As mentioned, speci…c input variables can be included

in equation (5.6) as along as technical ine¢ciency e¤ects are stochastic andinput variables in the production function are exogenous to the compositeerror term (Battese and Coelli 1995, and also, Forsund et al., 1980 andSchmidt and Lovell 1979)

Additional likelihood ratio (LR) tests are summarized in table 2 Correctcritical values from a mixed -squared distribution (at the 5 per cent level

of signi…cance) are drawn from Kodde and Palm (1986) The relevant teststatistic is

LR = 2fln[L(H0)=L(H1)]g = 2fln[L(H0)] ln[L(H1)]g (5.7)where L(H0) and L(H1) are the values of the likelihood function underthe null and alternative hypotheses respectively The null hypothesis of adeterministic time trend in equation (5.6) is rejected The null hypothesisthat technical ine¢ciency e¤ects are absent ( = 0 = 1 = 2 = 3 =

4= 0) and that farm-speci…c e¤ects do not in‡uence technical ine¢ciencies( 1 = 2 = 3 = 4 = 0) in equation (5.6) are both rejected, as is 0 = 1 =

2= 3 = 4 = 0 Finally, the null hypothesis that = 2

u=( 2

v+ 2

u) = 0, orthat ine¢ciency e¤ects are not stochastic, is rejected All results indicate thestochastic e¤ects and technical ine¢ciency matter and thus that traditionalOLS estimates are not appropriate in this study

Table 3 summarizes the results for the stochastic production frontier andine¢ciency models The coe¢cients on capital, labour, land and materialinputs are 0.17, 0.13, 0.24 and 0.51 respectively A time trend also tests assigni…cant at 1.1 per cent per year Results show that farms in the mainrice growing regions, those with larger farm size, and farms with a higherproportion of rice land ploughed by tractor are more e¢cient The size of thebinary variable SOIL is perhaps the least surprising Superior conditionsfor growing rice in the MRD and RRD, compared especially to the highlands

in the northwest or central areas (regions 3 and 6), are clearly re‡ected inprovincial-wide measures of e¢ciency throughout the sample period TheMRD in particular consistently ranks best in e¢ciency, year-to-year, and thee¢ciency measures for the MRD and RRD (taken together) are 11 to 13 percent higher throughout than the average for Vietnam as a whole (Detailedresults for each province and region by year are available from the authors

on request.) The policy requirement, in the past, that rice be produced inevery province of Vietnam, and the current practical restrictions on land use,

as detailed in section 2 above, thus appear unwarranted, at least in terms

of the potential loss in e¢ciency that results from producing rice outside ofthe Mekong and Red River Deltas

The coe¢cient for the proportion of rice land ploughed by tractor (T L)

is also substantial at -0.35, and remains large even when testing with theMRD and RRD taken separately An increase in number of tractors in rice

…elds clearly increases e¢ciency The are two policy concerns here First,and most importantly, the absence of credit markets and, in some cases, lessthan secure property rights, as discussed in section 2 above, undoubtedlylimits the amount of tractors in rice production Transactions costs on loans

in rural areas are prohibitive and when granted are often for terms of onlyone year or less Indeed, much of the increase in agricultural capital in thereform periods, and after, is due to accumulation from retained earnings, andnot from borrowing (see Che et al 2001 and Kompas 2004) Second, landpolicy itself often makes it di¢cult to employ tractors in rice …elds Plotsare often small and butt directly to adjoining plots (separated only by amound of dirt) and practical restrictions against farm size and impediments

to land consolidation that would help ensure contiguous or non-fragmentedplots (especially in the north), often make the use of tractors impractical,

or at least not without a good deal of cooperation among farmers

The coe¢cient of average farm size is smaller than might be expected,but still indicates that restrictions on farm size limit e¢ciency However, thisvalue rises considerably when estimating over the RRD and MRD (regions ofcomparable fertility) taken separately In these truncated data sets, the co-e¢cient on average farm size in the technical ine¢ciency model is -2.7 in theRRD, while in the MRD it is -0.1, both signi…cant at the one percent level.This is as expected In the RRD, where restrictions on farm size are moresevere and more broadly enforced, average farm size per crop is small at 0.4hectare per farm, compared to 1.4 hectares per farm, per crop, in the MRD,

so that e¢ciency gains are far from exhausted The reason for smaller farmsize in the RRD is usually attributed to a high population density in ruralareas in the north combined with explicit legal and moral restrictions against

‘excessive land accumulation’, at least in practice Moreover, although landcan be leased for up to 20 years, there still are only limited markets for theexchange of land or land-leases (GSO (VHLSS) 2004) Thus, smaller farmsize, the consequent smaller proportion of tractors used in rice …elds, morerestrictive land regulations and the slightly worse natural soil conditions inthe RRD explain the lower levels of e¢ciency compared to the MRD.The coe¢cient on the number of tractors, as opposed to the proportion ofrice land ploughed by tractors, is positive for the simple reason that in mostrural areas (other than the MRD and RRD) tractors are used for generaltransportation and for other industrial crops or small-scale industry Whentesting for the MRD and RRD alone, where tractors are largely dedicatedfor rice production, the coe¢cient tests larger, at -0.18, as expected.Finally, although average technical e¢ciency is low for Vietnam as awhole (59.2 percent) it is clear from Frontier 4.1 output that e¢ciency forrice farms in Vietnam and in the principal rice growing provinces (MRD and

RRD) has been rising over time, albeit slowly, from roughly 55 to 65 per cent

in Vietnam as a whole and 66 to 78 per cent for the principal rice growingareas The gradual increase in the amount of capital (tractors and bu¤alo) isundoubtedly one of the key explanations for this trend The only exception

is the year 1994 where all areas experienced a fall in e¢ciency and especially

so in the MRD and RRD The reason for this fall appears to be partly due toResolution 5 (Nguyen 1995), outlined …rst in 1993, which further redividedfarm size into smaller and non-contiguous plots, allocated now across priorfamily farm members, but perhaps mostly to the exceptional ‡oods in thatyear in most of the principal rice growing regions Program output showsthat previous technical e¢ciency measures were not recovered until three offour years later, or in 1997 for Vietnam as a whole and 1998 for the principalrice growing areas

The second frontier estimate uses survey data obtained from a random lection of 338 farms producing rice from 32 communes across 8 provinces inthe RRD and MRD, with a roughly equal split of farms and communes ineach area The survey was carried out from August to December 2004, withdetailed collection of all rice output and input data, as well as farm speci…ccharacteristics The main areas from which farms were selected in the MRDare Soc Trang, Tra Vinh, Vinh Long and Can Tho; and, in the RRD, from

se-Ha Tay, Nam Dinh, Thai Binh and Nam se-Ha Summary statistics are vided in table 4 Log likelihood ratio tests (available from the authors onrequest) con…rm the speci…cation given by

pro-ln Yi= 0+ 1ln Ki+ 2ln LABi+ 3ln LANi+ 4ln Fi (5.8)

+ 5ln Pi+ 6RRDi+ vi ui

for Y the output of paddy in kilograms, K capital in machinery hours, as thesum of hours a farm uses tractors in land preparation and transportation,pumps and threshing machines, LAB working days, as the sum of familyand hired labour, LAN total land size in hectares, F kilograms of fertilizerused, P pesticides in kilograms and RRD a binary variable for Red RiverDelta rice farms The ine¢ciency model in this case is

+ 4IRRi+ 5EDi+ !i

for SIZE the amount of cultivated rice land (both leased and directly trolled by the household), P LOT S the number of plots of rice land in agiven farm, as a proxy for land fragmentation and SOIL a measure of soilquality, ranked in decreasing order (from 1 to 6), based mainly on the chem-ical composition of the soil IRR is a measure of water availability (natural

con-and irrigated), ranked in decreasing order (from 1 to 4), obtained by askingfarmers to rank their irrigation conditions, based on the level and di¢culty

of supplying water and drainage The ranking from 1 to 4 is simply givenby: very good, good, fair or poor ED is the level of education of the farmdecision maker, categorized by four levels: primary, secondary, high schooland higher education

Results for the farm survey data set are reported in table 5 Given thenature of the data, the estimated input coe¢cients vary considerably withthe results from the provincial data set There are two reasons for this First,the provincial data set is nation-wide, with large variations in rice productionacross 60 provinces, and especially so compared to farm survey data in theprincipal rice growing regions Second, and perhaps more importantly, themeasure of inputs in each data set is vastly di¤erent For example, land inthe farm survey data refers to the actual value of land cultivated, rather than

a multiple of land cultivated over all rice crops during the course of a year,and capital is a value measure of all machines, rather than a constructedmeasure of bu¤aloes and tractors The value of the binary variable RRD inthe stochastic production, in table 5, alternatively, is straightforward andillustrates the advantages to growing rice in the south, compared to thenorth This value is -0.184 and is consistent with the measured di¤erence inTFP between the RRD and the MRD, illustrated in …gure 2

Of particular interest, however, are the ine¢ciency results Soil and gation are as expected, since they are ranked in decreasing order of quality,implying that higher quality soil and better irrigation increase e¢ciency,and it is clear that more educated farmers are also more e¢cient The co-e¢cients on SIZE and P LOT S indicate the loss in e¢ciency from currentland use practice, in a way that is not possible in the provincial data set,with provincial averages on farm size and no measure of plot numbers Theestimates clearly indicate that larger farms and farms with fewer plots aremore e¢cient The latter in particular indicates a potential issue with landfragmentation Admittedly, simply counting the number of plots in a givenfarm is a crude indicator of fragmentation, since it lacks a measure of dis-tance between plots or whether plots are contiguous or not, but it is alsoclear from the discussion in section 2 above that the more plots a farm has

irri-in Vietnam the more likely it is these plots are not contiguous This is cially so in the north, where, as indicted above, small and highly fragmentedfarms predominate Frontier estimates by Hung et al (2007), on a smallersurvey data set for 188 farms in the north only, near Hanoi in the RRD,

espe-in the year 2000, also show a negative relationship between the number ofplots and farm e¢ciency

5.6 Econometric speci…cation: VHLSS data (2004)

The third frontier estimate uses VHLSS data for 3,865 households in 2004largely engaged in rice production (from a total of more than 9000 house-holds surveyed) Although not literally a pure sample of rice producers,the rice output of the households in this sub-sample accounts for more than

75 per cent of total household annual crops in terms of quantity, and morethan 78 per cent in terms of value Summary statistics are listed in table 6.Log-likelihood ratio tests (available from the authors on request) generate aspeci…cation for the stochastic production frontier of the form

ln Yi = 0+ 1ln LANi+ 2ln LABi+ 3ln HLABi+ 4ln Mi (5.10)+ 5ln M Ri+ 6ln Fi+ 7ln Hi+ 8M RRDi+ vi ui

with an ine¢ciency model given by

+ 4QU ALi+ 5EXTi+ !i

for Y the output of paddy in kilograms, produced over the twelve monthsprior to the survey date, and LAN the amount of area (in square meters)that the household uses for annual crop production, regardless of its own-ership Labour comes from two sources: LAB household labour (in hours)and HLAB hired labour (in Vietnamese Dong) The values of machines(M ), (i.e., tractors, tools and implements), rented machines (M R), fertil-izer (F ) and herbicide (H) are all measured in constant-value VietnameseDong M RRD is a binary variable for rice produced in the MRD and RRD

In the ine¢ciency model, P LOT S is the number of separate annual cultural land plots in a household farm and ED is a rank for the education

agri-of the household head, given by numbers 0 to 5, or no schooling, primary,lower secondary, upper secondary, vocational training and college or uni-versity schooling CERT designates that a farm household holds a landcerti…cate title (measured as a ratio of land under title to total land size),allowing for the sale or lease of all or some plots of land and QU AL is ameasure of land quality, based on the land tax system, and generally corre-lated with the amount of soil nutrients and the proportion of soil serviced

by natural or irrigated water Annual agricultural land is classi…ed into 6categories which serve as the basis for the government to collect agriculturaltaxes In equation (5.11), QU AL is speci…cally the ratio of the annual agri-cultural land area of the best two land types over total land holdings EXT

is a binary variable simply measured by a visit to an extension services

of-…ce, attending meetings to seek advice or guidance on cultivation practices

or raising livestock, or by being visited on farm by an extension sta¤ o¢cer

5.7 Results for VHLSS data

Results for the VHLSS data set are reported in table 7 Estimated inputcoe¢cients are comparable to the results for the farm survey data set Thebinary variable M RRD indicates the advantages of growing rice in the maindelta areas (A alternative speci…cation, with M RRD in the technical in-e¢ciency model, as in the estimates using the provincial data set above,generates similar results.) Results again indicate that increases in the num-ber of plots (as a proxy for land fragmentation), decrease e¢ciency, andalso that better educated farmers and higher quality soil (in terms of wateravailability and irrigation) increase e¢ciency across farms In the VHLSSdata land quality varies considerably and the mean is low (see table 6), in-dicating that rice is produced in many areas without the natural advantage

of water availability or irrigation This is in sharp contrast to the resultsfor the farm survey data set above, drawn mostly from farms in the MRDand RRD ,where water is not as much of an issue, and average land quality

by this measure is much higher (see table 4) Of added interest here arecoe¢cient estimates on land use certi…cate and access to extension services

As mentioned, a proper land use certi…cate is essential not only for the ease

of acquiring, selling or leasing land, but it also provides the often only readysource of collateral for farm loans Those farms with a proper certi…cateare more e¢cient, as are those (nearly half the sample) that have access toagricultural extension services

6 Closing Remarks

Extensive land and market reform in Vietnam has resulted in dramatic creases in rice output over the past thirty years Results show that TFPincreases considerably in the major rice growing areas (the Mekong and RedRiver Delta areas) during the early years of land and market reform, butwith clear evidence of a productivity ‘slow-down’ since 2000 in all regionsexcept the MRD TFP in the MRD remains much higher than in the RRD,and TFP in other regions (and especially in poor areas) remains virtuallyunchanged throughout the entire period Terms of trade and net returns arealso favorable throughout the reform period, providing much of the explana-tion for increased incomes and poverty reduction during this time, but over-all performance has worsened considerably since 2001 The di¤erences overtime and by region speak directly to existing land use regulations and prac-tices and suggest calls for further land and market reform In this regard,additional frontier and e¢ciency model estimates illustrate the remaininginstitutional and policy constraints, including existing restrictions on landconsolidation and conversion and poorly developed markets for land andcapital Estimates show that larger and less land-fragmented farms, farms

in-in the major rice growin-ing areas, and those farms that are better irrigated,

have a greater proportion of capital per unit of cultivated land, a clear erty right or land use certi…cate and access to agricultural extension servicesare more e¢cient.

prop-With this in mind, it seems clear that growing rice in every province, atleast terms of a narrowly de…ned e¢ciency criteria, is inappropriate Pro-ductivity and e¢ciency are both substantially larger in the Mekong and RedRiver delta areas, where rice production has a clear comparative advantage.This shows up repeatedly in both TFP and related measures, as well as infrontier and ine¢ciency models, in terms of both the magnitude of binaryvariables for these regions (and their e¤ects on output or e¢ciency), and co-e¢cient estimates that measure the e¤ects of irrigation or water availability

on e¢ciency Land policy (formal or in practice) which makes it di¢cult forland to be converted to other uses thus can not be justi…ed on these grounds.The same can be said for land consolidation If farms that are larger and lessfragmented are more e¢cient, practical restrictions on land size needs to berelaxed and a more active real estate market for land needs to be provided,encouraging low-cost and e¢ciency enhancing land transfers A necessaryand straightforward pre-requisite for this is well-de…ned land use certi…cates,covering every parcel of land, something that Vietnam has yet not been able

to accomplish This may also partly resolve problems with credit ity, as would a signi…cant extension of the 20 year lease provisions on parcels

availabil-of agricultural land The current leases on land allocated in 1993 are indeedabout to expire Without a land use certi…cate, or with limited remainingtenure, it is di¢cult if not impossible to secure a loan, much less convertand consolidate land The original land and market reforms, as dramatic

as they were, have not gone far enough to secure property rights or providesu¢cient or suitable markets for land and capital

There are at least three issues are warrant further research First, itwould be useful to have a better de…ned measure of land fragmentation thanused here, one that includes distance and a spatial representation of non-contiguous plots Although the number of plots in a given farm is a usefulproxy, and perhaps more than su¢cient for Vietnam, a more re…ned indexwould be useful This may also further clarify any potential interactione¤ects between fragmentation and land size on ine¢ciency Second, theestimates would bene…t from additional measures of rural services Theonly variable used here, access to agricultural extension services, as a simplebinary variable, matters greatly to e¢ciency, but so too must variables likerural infrastructure (e.g., roads, water rights and quarantine and surveillancemeasures) and speci…c cultivation practices, including the use of rice hybrids.Unfortunately, there is a lack of broad rice farm survey data to providesuch estimates Finally, and perhaps most importantly, there needs to be aclear investigation into the precise nature and cause of the thin or poorlydeveloped agricultural land and credit markets in Vietnam, and what speci…cpolicies might be best to help resolve these constraints It is undoubtedly