Professional Documents
Culture Documents
Reddaway and Md Mizanur Rahman Reviewed work(s): Source: Economic and Political Weekly, Vol. 11, No. 23 (Jun. 5, 1976), pp. 843-849 Published by: Economic and Political Weekly Stable URL: http://www.jstor.org/stable/4364685 . Accessed: 26/11/2011 04:20
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Economic and Political Weekly is collaborating with JSTOR to digitize, preserve and extend access to Economic and Political Weekly.
http://www.jstor.org
SPECIALARTICLES
The
Scale
of
Smuggling
Out
of
Bangladesh
I
Rice
PRINCIPLES OF METHOD USED FOR RICE
Inevitably, no statistics exist about the scale of smuggling operations. The approach in this Note is to seek some indirect guidance from price statistics which have already been collected for different parts of the country. Since we wrote this paper, we have learned that an intelligent collection of local gossip in border areas led to conclusions broadly similar to ours. The principle of the method is simple. Suppose that the price of a standard quality of rice is known for each of a number of rural centres near the border, and that these are all "surplus areas" i e, they produce more rice than is consumed locally. In the absence of smuggling, the surplus would be transported towards the deficit areas of Bangladesh, notably Dacca. Suppose that prices are also known for the same quality of rice in rural centres at distances of? (say) 3 and 10 miles inland from each border place and that these are on the route which the surplus rice would normally follow in seeking an internal market; suppose furTABrEA Category (1) (2) (3) (4) (5) (6) (7) (8) (9) Value in Thousand Takas Gold/Silver 481 Fish 349 Currency 206 Medicine 182 Eggs and milk powder 154 Gunny bags 126 Rice/Paddy 114 Flour 106 Skins 90
(UO.jue
843
June 5, 1976
TABLEI Border Place Map Code Name Distance* from Frontier
R1 R2 R3
R4 R5
3 6 7 9 11 11 19
**
16 31 22 32 45 24 22 20
R6
R7
R8
* **
Distances (in miles) are measured straight, without regard to means of transport. Smuggling would be by boat: Barguna is on a navigable river. near the coast.
(t) Two possible explanations for the relatively high price at the border in March are available, apart from erratic price reporting:
(i) Smuggling was rampant in Mlrch, but was greatly reduced by the Army's arrival in April. (ii) The border area had run short of rice in March. but got fresh supplies in May from the boro crop. ther that these centres are also surplus areas. The following points can be made: (a) In the absence of smuggling, the prices recorded would show an upward progression as one went inland from the border, to reflect the cost of transport. (b) If smuggling were a really major factor, then rice would fetch a higher price at the border place than in the two neighbouring inland centres, so as to attract supplies to the frontier. (c) Intermediately, if rice were smuggled out on a small scale, so as not even to absorb the surplus of the border areas, then the price would still be lowest at the border and rise progressively (but less rice would move inland from the border area than in our first case). (d) If rice were smuggled out on a moderate scale, so as to absorb the whole of the surplus in the border area to a depth of about 5 miles from the frontier, then the price would fall between the border centre and the one 3 miles inland, and then rise in the one 10 miles inland. Inverting this collection of statements and making some simplifying assumptions, one would be able to say (if all data were available and perfectly accurate) that the place where the price was
844
lowest would miarkthe cut-off point. Smuggling would absorb the surpluses (not the whole production) of the area between that point and the border. This conclusion would be upset if smuggling were largely done by big operators who loaded lorries at places far inland, and then drove them over the border. This procedure seems rather improbable, as the operators could save money by buying and loading the rice nearer to the border. We gather moreover that the "local gossip" mentioned above lent no support to the idea that such operations were quantitatively important.
DATA PROBLEMS FOR RICE
Inevitably,
data
dure which one would like to follow. In the case of rice, the free market price for coarse rice is reported daily to the Ministry of Food and Civil Supplies in respect of some 80 places. Our first step was, therefore, to look at all of these which could be reasonably regarded as "near the border", and to see whether there was a second place on the list which could be treated as the "inland neighbour", i e, which was reasonably near, and preferably on the natural route by which supplies would go to internal markets. In a very few cases there was a third place further inland which could be regarded as meeting the requirements of the "10 miles away centre" mentioned in the previous section: these were however so rare that we could make no systematic use of them.
OF JUTE NEAR FRON1IER AND INLAND
2:
COMPARISON
BEIWEEN
PRICES
[See map] Map Code Neighbour place Border Place Distance' Distance* Name from from Na-me
Frontier Border
Place
Price in Neighbour as Per Cent of Price ill Border Sept 74 Sept 73 98.6 104.3 101.0 95.9 105.2 98.5 98.3 103.7
J1
J2 J3 J.
J5
*
101.0 87.4 15 9 Rupdia miles) are measured straight, without regard to means of Distances (in transport.
3 4 7 9
16 24 22 9
50.7 11.6
20.0 7.5
one day different.
50.7 11.6
20.0 7.5
Given these pairs, it was tempting to look separately at the figures for each of them, to see whether the price at the border was higher: if so, this would seem to suggest smuggling on a scale large enough to attract rice from inland to that border place. Unfortunately, however, inspection of the figures made it clear that they were not adequate to serve as a basis for such a simple operation. Thus, (a) The prices in each place showed erratic movements from one day to the next, so that the relative prices of neighbour and border might change erratically. The best way of treating them seemed to be to take an average of a good nuimber of days in the month, to represent that rmonth, and hope that errorsof reporting, quality variations, etc, would then largely average out. (b) It is clearly possible that, when rice is traded at various prices in each centre and varies in quality even within the category "coarse rice", the methods adopted by the reporter in the border place may vary from those used in the neighbouiringplace and introduce an error into the comparison. Consequently the most that one can really do with any safety is to provide material for an overall assessment, rather than arrive at conclusions for each area separately. And by repeating the calculations for three different months, one can provide some additional guidance which is particularly useful if the conclusions from each month's figures are much the same. One must however recognise that the reporter at (say) Rajshahi might consistently tend to use methods- which led him to report unduly,h4gh (or low) in relation to the
reporter at Natore, so that the data for the different months are not completely
independent.
RESULTS FOR RICE
price in the inland (or "neighbour") place as a percentage of the price in the border place, for each of the five months. Looking at these figures as a group, there is not much doubt about the main impression which they create. Only twelve out of the 40 figures are below 100 (ignoring one of 99.9), and no pair of places has more than two months below 100: this suggests the generalisation that smuggling was typically on a scale not big enough to attract rice from the inland place to the border. Moreover, most of the figures which are below 100 would be readily explicable on the hypothesis that errors in the data would be likely to produce a few perverse figures: possible explanations of the two figures below 90 are given under the Table, but it is doubtful whether one should attempt to explain intheir basis is so dividual figures wNhen uncertain.2
COMMrENTrS ON RESULTS FOR RICE
We thought it hest to work with prices for months befor? the 1974 flood, since this might well have introduced additional disturbances in the reporting system; as well as producing some strange situations. We therefore used prices for December 1973, January 1974, February 1974, March 1974 and May 1974: in all these months except May, smuggling was alleged to be on a substantial scale, and was the subject of miuch discussion. It is also of some interest to note that the army started patrolling the border in April, which we thought might make a comparison between, March and May of particular interest. The pairs of places which we used (which include a number which very dubiously met-our requirem,ents,since there were so few good cases) are shown in Table 1. This Tfable also shows the
TABLE
If our border places had all been really near the frontier, and if the neighbour places had been within 10 miles of them, the above result would have been a very important one: it would mean that smuggling, was, at most, taking the surpluses (not the whole output) produced in the area within (say) 5 miles of the frontier. Unfortunately, the lack of suitable reporting centres forced us to take as border places ones which were further from. the frontier than we would have wished (see Table 1). Even if we could establish that all our border-places wvere outside the area from which rice was smuggled, there would still be a moderately suibstantial area nearer the border about which we could say nothing. Formally, our results wotuld be consistent with the, be(INCLUDING MESHTA) IN BANGLADESH
4:
SUPPLY
AND USAGE
OF JUTE
(lakh bales)
1972/73 Supply 1973/74
End-season stock* Apparent Usage Usage Exports Mill Consumption Growers' Consumption (estimated) Destroyed and unaccounted for Apparent Usage
*
une
,
TABLE
lief that there is no smuggling at all; but equally they woud be consistent with the belief that it absorbs the whole surplus of that xarea. Worse still, the neighbour places are on average about 25 miles away, so that we are testinig whether an allegedly high demand from smugglers is sufficiently powerful to divert rice from quite a distance inland to the border places. At the risk of straining the data, it seems worthwyhileto try to set an upper limit to the likely scale of smuggling. We therefore assume, rather arbitrarily, that a "typical" border place (i e, one at the average distance from the frontier) is within the area seriously affected by smuggling, and try to assess how much furthc-rinland the area might stretch. As the "neighbour" price was on average higher (by 3.1 per cent), the cut-off point would seem to be typically nearer to the border place than to the neighbour. This would mean that the area from which the surplus might be regarded as potentially available for smuggling would extend inwards from the frontier to a depth equal to the distance to the border place (averaging about 10 miles) plus perhaps one-quarter of the distance between the border place and the neighbour,3 making a total of about 15 to 20 miles. The distance, we repeat, is meant to give an upper limit to the area from which the surplus might be regarded as available for smuggling. As a very rough test of this conclusion, we tried eliminating the three pairs of places where thnedistance between them exceeds 25 miles, and the result is, therefore, most uncertain. This leaves the broad picture little affected. Of the 25 rem,aining figures in Table 1, eight are below 100 (i e, again rather under one-third); on average the neighbour price is again higher, by 2.7 per cent this time instead of .3.1 per cent; the average distance of the border place from the frontier is 10 miles; and the final conclusion would be slightly reduced to about 15 miles. We do not feel competent to translate this concluision into a figure for the maximum tonnage of rice smuggled per vear, because we do not know how to assess the surplus of rice in a border area with a depth of about 15 miles. We do wish to emphasise however the arithmetic point that the area in question cannot be assessed by taking the full length of the boundary,with all its wiggles and curves, and 846
1973/74
Exports Mill Consumption Domestic Consumption Apparent Usage mutiplying by 15: this counts most places many times over, because they are within 15 miles of more than one spot on the boundary. The relevant area may perhaps he about one-sixth of the area of Bangladesh.4
ed that it was caused by large-scale smuggling of rice out of the country. On general principles, however, it is equally possible to argue that the high price was caused by production plus imports being inadequate to meet the demand: the high price would then AN Al.TERNATIVE APPROACH make smuggling unprofitable. The test We intended to apply an alternative between these two theories is to assess approach (which is described more fully the gross profit on smujgglingby using in the section on jute), based on a the Indian price and the black-market direct comparison of rice prices on exchange rate, and to supplemenit this either side of the border, using the direct assessment by the calculations black-market rate of exchange. The set out in the Jute section. We feel only relevant data Nvhichwe were able that people with better access to data to secure, however, were rice prices in should be able to apply the test much Comilla and Agartala, and the black- more conclusively than we could: our market rate of exchange on the border own limited efforts do however support between them. These figures, which ran the results of our main approach by from August 1974 to November 1974, favouring the small smuggling hypoconsistently showed that the Indian thesis. price was substantially lowuer than the II Bangladesh price, even when converted at the black-market rate of exchange; Jute there was, in consequence, no apparent In the case of jute, we first applied incentive to smuggle rice from Comilla to Agartala. Moreover, it is very doubt- the same method as for rice. The data ful whether this result can be explained available are howveverboth less abunaway in terms of non-comparabilities in dant and not so suitable to our purthe basis on which the prices were re- pose. Thus the number of places fronm the converted Indian which prices are collected by the Direccorded, since price -as typically some 20 per cent torate of Agricultural Marketing is smaller; prices are collected weekly, inbelow the Bangladesh price. It does not seem to us surprising stead of daily, and are quite often not that the Bangladesh price of rice, in a reported (possibly because there is inperiod of near-famine, should be so suflicient business); and the designation high as to make it unprofitableto smug- of what is covered (loose unassorted gle rice out of the country. The black- jute) makes it clear that there may be market rate for the taka was of course serious variations in quality, both betproviding a stimulus to smuggle some- ween different places and between two thing out, but that something need not dates in the same place. In consequence, we could only estabe rice (as explained in the first section). This point applies whatever the blish five usable pairs of places, and aggregate amount of smuggling may be. Bhairab-bazar acts as "neighbour" to It may help to conclude this section two different border places. The disby explaining a possible source of con- tances involved are however shorter, fusion abouit the probable scale of which improves the sensitivity of the smuggling. On observing the unprece- test. We took September as the month dentedly high price of rice in Bangladesh, many people seem to have inferr- when supplies of jute are likely to be
June 5, 1976
important Indian jute centre, situated just over the' eastern border of Bangladesh, and close to two of our "border places" Companyganj and Brahinanbaria. Furthermore, we also obtained the black-market' rate of exchange between rupees and takas which prevailed at that part of the frontier. This enabled us to calculate the apparent gross profit obtainable on smluggling jute from Companyganj to Agartala, by converting the Indian price to takas at the black-market rate of exchange, and subtracting the price at Companyganj; and similarly for Brahmanbaria. The figures for September 7, 1974 (the first date for which they are available) show the calculation, and are of some interest in themselves: Black-marketvalue of 100 takas 44 rupees Price of jute at Agartala, per maund 39 rupees Equivalent in takas 88.6 takas Price of jute at Companyganj, per maund 74.0 takas Gross profit on smuggling, per maund 14.6 takas This calculation involves the danger that the quotations in Agartala and Companyganj might be on a different basis, and of course the black-market rate which was reported to us may be biased (upwards or downwards) compared with the one likely to be obtained by the actual smugglers. If, however, one does the calculations for a series of dates, these biases are likely to be similar oni each occasion, so that one gets quite a good idea of whether the smuggling profit i's or is not fairly constant, and which dates showed a high (or low) profit. Now if smuggling is a rcally major factor, the profit should not vary very much: a rise in Indian demand and in the Indian price will drive up the price at Companygunj, and this will attract large supplies from inside Bangladesh by establishing a price in Companyganj which is above the inland prices. If, on the other hand, smuggling takes only a small fraction of the local output from, the Bangladesh side of the border, and cannot rapidly expand to take advantage of the high price at Agartala, then the smuggling profit can vary a good deal, and there will. be no tendency for a hiigh smiugglingprofit to raise the Companyganj price above the itnland price. This gives us, then, a double test as to which hypothesis fits the facts best: (1) "Big smuggling" predicts fairly stable smuggling profits. "Small
847
bomp&A-
scale
20
miles
(INDIA)
R3J3
QR3 J3
R6 R5
R2. R7 0
Q
~~~~R2
(INDIA)
R6
0
(INDIA)
R5
B Basar J2J14
Q 0
Rl Ji
J4
ij o El Ll J5 ~R4
J2
GanoC
o UA
?R4
~~~~R
CAC I
OFP BAY BENGAL boundary
capital
o3
International
best, and we covered both 1973 and 1974. The results, for what they are worth, are shown in Table 2. This shows that each pair of places has one figure below 100 and one above; and with one exception, all the figures are fairly close to 100. In trying to draw conclusions, one must remember that jute differs from rice in one very important respect. There is virtually no local consumption of jute, so that the whole output of a border place might be smuggled out without attracting supplies from inland: this is in marked contrast to rice, where only the surplus over local consumption was available for smuggling or inland marketing.
It seems, therefore, that Table 2 presents little or no evidence in favour of the hypothesis that more jute moves towards the frontier for smuiggling, rather than toxvards the mills of Dacca and Khulna and the legitimate exporters, but one cannot carry the analysis much further.
AN ALTERNATIVE APPIROACH We obtained a further clue by a
method which we had hoped to use on a larger scale. The available data, however, allowed only a limited use of the method, but one which is better than
nothing.
June 5 1976
smuggling" makes no prediction: Batik in its Selected Economic Indismuggling profifs could quite cators - appear to suggest that smugwell be very variable. gling must be very smalL Table 4 (2) "Big smuggling" predicts that shows the position for -1972-73 and when smuggling profits are high, 1973-74. the price at Companyganj will Smuggling is part of the item for exceed that at Bhairab-bazar. "Small smugglinig" makes no "destroyed and unaccounted for" and such prediction: the price at on the face of it, the Table sets a very Bhairab-bazar xvill probably con- low upper limit, since the losses by tinue to he higher, since some fire were reported as 1.92 lakh bales and 1.80 lakh bales in the two years: jute moves that way. even if these claims exaggerated the Table 3 provides all the data which amount actually destroyed, smuggling we could obtain to apply these tests. could hardly amount to more than The smuggling profit for each date is about 1 per cent of the crop, unless set against what we call the "legal trade there are serious errors in some of the profit", i e, the difference between the other items. Bhairab-bazar price and the price at the border, which is the reward for Of these other items, growers' conmoving jute towards the mills and other sumption is only an estimate, but it legitimate buyers. cannot be negative so that there is There can be no doubt about what little scope for raising the residual figure the Table tells us: by reducing it. The stock figures (1) The (gross) smuggling profit is similarly include an estimate for stocks very variable indeed, even within held by growers and petty traders, but this 9-week period. It twice rose at the end of June, (before the harvest) to over 50 takas per maund and sank very low in the second half this: figure is bound to be very small of September (being actually and the difference between the opening negative on 28th September for and closing stock is almost negligible. Companygani. This is certainly not what the big smuggling hy; The crucial question, therefore, is pothesis predicted. how far the figure given for production (2) When the smuggling profit is high, the legal trade profit does is to be regarded as an independent not become negative (as the big figure - based on the acreage devoted smuggling hypothesis predicts): to, jute and an average yield derived on the contrary it became very high itself. The rank correlation from a sample, as in the case of rice for the two series, is not far production - and how far it is derived short of + 1. instead of being from the usage figures by assuming a negative. figure for "destroyed and unaccounted One must not of course exaggerate for". If it is wholly the latter, then the importance of this supplementary the Table shows only that the people evidence, which is based on a very who produced the production figure limited amount of data. What one should assume smuggling to be small. This is have used all say is simply that we quite significant evidence in itself the data which we could find,-and that they are closely concerned in the it consistently showed results entirely matter - but it would be more helpful different from those predicted by the to work from an independent estimate big smuggling hypothesis. based on returns from the fields.5
SECOND ALTERNATIVE APPROACH THIRD ALTERNAUVE APPROACH
The second and third alternative approaches should be capable of giving two good, and independent, indications of the level of smuggling; people with fuller access to the statistical sources should be able to arrive at a fairly clear view. It would however obviously requi.re a radical alteration of the production, figures (upwards in Bangladesh, downwards in India) to provide evidence in favour of the "big smuggling" hypothesis (taken as meaning that 10 per cent or more of the crop was smuggled into India). Taking these statistical approaches in conjunction with our own indirect approaches, we feel bound to express scepticism about the big smuggling hypothesis for jute, just as we did for rice. It seems to us that the quantitative importance of smuggling has been exaggerated, with a resultant detrimental effect on assessments of how well the country has been governed. Needless to say, this does not imply hat efforts to cut down smuggling should be relaxed.
In the case of jute (but not of rice), It should also be possible to get an idea of the possible amount smuggled by combining the figures for the main types of usage (which are known fairly accurately) with the figures for production, plus any rundown in stocks. There are some items in the comparison which have to be estimated, but they are usually known to be small, so that errors in them should not seriously affect the assessment. Figures produced by the Jute Board also published by the Bangladesh 848
Another way of making an estimate is to use the Indian figures, and to apply the same technique as in the second approach - but looking , this time a residual on the supply side, to reflect smuggling into the country. From the Monthly Summary of the Indian Jute Mills Association, we were able to derive the figures shown in Table 5. The 1972/73 figures balance exactly, without any room for smuggling., h)utthey raise again the question of howv the prodluction figure is derived.
Perhaps we should explain that we also extracted prices for Dinajpur, as "border", and Thaknurgaon, as "inland laeighbour"', and these
ECONOMIICAND POLITICAL WVEEKLY showed figures around 90 for all five months. We, then realised, however, that our preoccupation with nearness to the border had led us to ignore the need for the "neighbour" to be on the route to the inland markets. Inspection of the map showed that the surplus rice from Thakurgaon would go to inland markets via Dinajpur, so that the price there would be higher, whether or not there was large-scale smuggling: nothing whatever can be deduced from the figure of 90, and we should not have included this pair of places on our original list. 3 The fraction is put low, because one really wants only the additionfl
June 5, 1976
distance away from the fron-
tier.: 4 If, to take purely illustrative figures, one assumed that this area produced one-fifth of the *country's rice, and that one-tenth of that output was surplus to the area's need for consumption and seed, and was all smuggled into India, then one comes to' rather over 2 lakh tons for what is intended to be a high estimate. This is not more than an illustration, but we feel confident that any figure like a million tons gets no support wrhatever from our results. 5 So far as we understand it, an independent estimate was so made and agreed closely for each of the two years.
Other things remaining the same, the supply of money is a highly stable, increasing function of adjustcd highpowered money. Perhaps, much of Mujurndar'sconfusion regarding the moneymultiplier theory could have been avoided had he read the first proposition in these terms. Let us illustrate this point. After summarising the main ingredients of the money-multiplier theory, he speaks of a corollary of the multiplier approach, viz, "the larger the proportion of bank reserves in reserve money, the higher the value of the money multiplier is likely tc be".5 Observing what he calls "the typical illustration of 1973-74", he considers the corollary misleading. No doubt, the' corollary is misleading. But the misleading nature of the corollary is inherited from the loosely-worded first proposition of the multiplier theory. A careful reading of the first proposition (as indicated above) would have led him to believe -that, through the technique of variaticns in the statutory reserve-ratio, the central bank can affect the surplW of adjusted high-powered money, leaving the money multiplier to be determined endogenously, by the behavicur of the banks and the public. Perhaps one could attribute this confusion about the money-mlultiplier theory to Gupta for not introducing the concept of adjusted high-powered money and its importance in the semantic exercise of section I of his paper. Later, when presenting the moneymultiplier theory also, he says: "In its simplest form, the theory says that the supply of money (MS) is a highly stable increasing function of high-powered money (H) alone. In other words, it says that as H changes M also changes in the same direction." 6 On another occasion too, he wrote: "Ordinarily, the higher the proportion of reserves in H, the greater also the high-poweredness of H in that the same H, other things being the same, will come to be associated with a larger amount of mnoney."7 One does not know whether Gupta included the constancy of the required reserve ratio in his "other things being the same" assumption. These misleading wordings of the money-multiplier theory might have caused confusion in the minds of many, including Mujumdar. But the fact that Gupta was very much aware of the limitations of such loose wording is apparent from one of his DSE working Papers in which he discusses this problem of adjusting the high-powered 849