How to view Structural Change: The Case of Economic Transition in Bulgaria

How to view Structural Change: The Case of
Economic Transition in Bulgaria
Gancho Ganchev, Lothar Krempel, Margarita Shivergeva∗
31st July 2001
The complex transformations, taking place in the post-communist countries
are hard to scrutinize and appraise theoretically. One of the ways to summarize the
changes is the input-output analysis of inter sectoral flows of a national economy.
However, the possibility of understanding the final outcome of all alterations in
the tables of inter sector flows is hampered by the extreme complexity of the
data. In addition, even if we were able to immediately process all the numbers
in the schedule, the question remains how to interpret them. A promising new
method to immediately grasp both qualitative dimensions and theoretical meaning
of complex data sets, are methods for the visualization of structures.
The latter is different from the usual in economical chart drawing. The visu-
alization is a class of techniques that allows for multidimensional representation
of manifold data series. The multiple dimensions effect is achieved by using fea-
tures like colors, distances, forms, shapes, grouping of objects etc. .In the case of
transition-generated economic evolution in Bulgaria, we used inter sector input-
output flows evaluated in dollars. Our aim was not simply to illustrate the flows
at different points of time, but to detect the emerging of a new, market-based eco-
nomic order.
It is apparent, that the core of the economic reforms in the former centrally
planned economies is the price liberalization. The decentralized relative price
changes lead to direct shifts of the inter sectoral flows, that can be easily grasped,
and to more complex indirect adjustments. According to economic theory, the
price changes should lead to an economic equilibrium if some general suppo-
sitions are fulfilled. These conjectures usually comprise the so called hard bud-
get constraints,Walras’ Law, and further homogeneity and behavioral assumptions
∗ Research Project at the Max-Planck Institute for the Study of Societies: Visualizing Economic
Tansition in Bulgaria. Project members are: Gantcho Ganchev, SASE BG , Sofia, Bulgaria (email:
gtganch@yahoo.com), Lothar Krempel, Max Planck Institute for the Study of Societies, Cologne,
Germany (email: krempel@mpi-fg-koeln.mpg.de), Margarita Shivergeva, Department of Political
Science, New Bulgarian University, Sofia, Bulgaria (email: mshivergeva@yahoo.com).
1

---

2
BG 1990 input-output use table
Forestry
flows exceeding 15 %
in $ US
Communication
Buildmat
Leather
Construction
Other_Ind
Wood
Agriculture
Transport
Textile
W_Apparel
Food
ChemOil
Coal_Oil_Gas
Engineering
Electr_Power
TradeFinTour
MetalOres
Printing
< low technology
< medium technology
< high technology
Woodpulp
Lothar Krempel, Max Planck Institute for the Study of Societies, Cologne
http://www.mpi−fg−koeln.mpg.de/~lk/netvis.html
©
5/02/01: 3:12:44−> W96B__.EPS Bg 96 select lines gt: 100
Figure 1: The Bulgarian Economy in 1990,
in the last year of the command economy period, when both national central
planning institutions and COMECON were still functioning.
like gross substitutability. If we assume that these equilibrium convergence con-
ditions are met, we can view the equilibrium as a balance between attractive and
repulsive forces. The drawing forces can be associated with the inter sectoral
exchange. The more intensive the exchange among sectors, the stronger the at-
tractive coercion amid them. Since the trade between sectors is associated with
attraction, the gross substitutability and the Walras’ Law are the natural candidates
for representation of the repulsive potential. Contrary to input-output flows, the
gross substitutability and Walras’ Law can be understood as system characteristics
or as homogeneity assumptions, which characterize the elements of the system.
More precisely, the system is supposed to react to price increases (decreases) in

---

3
any sector by an increase (decrease) of the excess demand in the other sectors,
while preserving zero summary excess demand in the economy as a whole.
BG 1996 input-output use table
Forestry
flows exceeding 15 %
in $ US
Woodpulp
Buildmat
Printing
Food
Construction
Food
Transport
TradeFinTour
Engineering
Agriculture
Wood
ChemOil
Leather
MetalOres
Communication
Coal_Oil_Gas
Electr_Power
Textile
Other_Ind
< low technology
< medium technology
< high technology
W_Apparel
Lothar Krempel, Max Planck Institute for the Study of Societies, Cologne
http://www.mpi−fg−koeln.mpg.de/~lk/netvis.html
©
5/02/01: 3:12:44−> W96B__.EPS Bg 96 select lines gt: 100
Figure 2: The Bulgarian Economy in 1996
after the shock therapy of 1991 and subsequent market oriented evolution.
Obviously the more important a given sector is as supplier and user, the stronger
the repulsive forces in terms of amore intensive summary reaction of the rest of
the economy to sector’s output price changes. This means that the variations of at-
tractive forces are matched by countervailing alterations of repulsive potential. In
equilibrium attractive and repelling forces cancel each other out what corresponds
to zero excess demand in economic terms. The key role of the gross substitutabil-
ity is evident. If the sectors were not gross substitutes but gross complements then
any deviation from the equilibrium would generate even higher aberration.
Surprisingly, these general ideas can be visualized. For this purpose we can

---

4
use the so called "spring embedders". A visualization technique allowing for com-
putations of equilibrium positions among objects with different attractive and re-
pulsive forces.
The general idea of these procedures is to understand an input-output matrix
as a valued graph in which economic sectors are linked by money flows and to
apply graph drawing procedures to produce layouts for the nodes. This locates
the sectors of the input-output graph in specific positions of a solution space.
Spring embedders are among various other graph-drawing techniques a very
flexible family of algorithms. They treat the flows between sectors as attractive
forces which are counterbalanced by repulsive electrical fields of a given size
which repulse all nodes from each other. Depending on the size of the repulsive
forces the layouts can be finetuned (spread or shrunk) while the neighborhoods
are maintained which enhances the overall readability.
The resulting layouts are system equilibria at low levels of potential energy
where all attractive and repulsive forces balance. Strongly connected nodes are
placed near while weakly connected units are placed distant to each other. Nodes
with strong flows to many other sectors move to the center of the system whereas
sectors are repulsed to the total systems’ periphery when they have few or weak
links only.
These layouts are often surprisingly easily to read, and can be optimized so
that the spacing of the nodes permit communication of additional information
about the flows and the sizes of the nodes. Enhanced images allow one to read
the position of a given sector in the overall structure of exchange and supply the
viewer with additional information: the volume of sectoral transactions are com-
municated by using arrows of different size and the size of the node symbols of a
sector represent the total of all flows a sector is involved in.
Such drawings can be further enriched if additional external data (attributes)
are available for the system. In such a case it is possible to map this external
information with colorschemes onto the layout. This can help to identify local
concentrations of such attributes for specific positions.
If we assume, that to any effective economic equilibrium state corresponds to
a computed space equilibrium, then we can regard the visual representations of the
latter as isomorphic images of the former. Since the equilibrium of the underlying
economy is assumed by definition, the visualization is a "neoclassical" one.The
results of the application of spring embedding procedure to the input-output data
of Bulgarian economy are presented at Figure 1 and 2.
Figure 1 first is based on 1990 data, the last year of command economy period
when both national central planning institutions and COMECON were still func-
tioning. The second makes use of 1996 figures, after the shock therapy of 1991
and subsequent market oriented evolution.
Inasmuch as our aim is to feature the transition shifts, it would be appropriate

---

5
to start with sectors that enter or leave the domain with the strongest attractive
forces (intensive input-output exchange). Sectors that are repulsed to the periph-
ery are first of all the so called other industries, as well as the wearing-apparel
industry. The possible explanation is the increased foreign competition, general
economic decline and the less intensive use of inputs from the textile sector in the
case of clothing.
On the other hand, we have two opposite examples - communications and
printing. Both industries were repelled by the command structures, but integrated
by the market forces. The latter happened because of the high importance of
branches, involved in information diffusion in a decentralized exchange economy.
There are also relatively stable configurations, like the connection construction-
building materials industry or the group of the heavy industries- extraction of coal,
oil, gas and metal ores, metallurgy, chemicals, electricity and engineering. The
volume and the connections of transportation also remained virtually unchanged.
However even within these relatively rigid arrangements we observe important
internal changes- the role of oil & gas as well as of chemicals increased, whereas,
the importance of engineering declined.
Yet, the most important mutation occurred in the tertiary sector. As we can
notice by comparing the Figures 1 and 2, the volume, the connections and the
positioning of the aggregated sector of trade, finance and tourism, transformed
dramatically. From a business, which was repelled by the command regime to
the economic periphery, the sector supplying exchange-related services became
central to the economy. Excluding the primary sectors like coal, oil, gas and
ores extraction, all the other branches are in fact integrated by the commerce and
finance activities. The increase of the value of services’ sector supplies and inputs
is not as spectacular as the change of its’ position in the economy. It clearly plays
the role of the organizing center.
This is in striking contrast with the central planning picture, where no such
central sector exists. This unexpected feature illustrates the fact, that the planning
center (State Planning Committee) of the command economy period was not an
economic sector, participating in the exchange, but a political institution of the
ruling party. Thus its’ directives were in a sense external shocks to the economy.
Contrary to command economy, the co-ordination in a decentralized market econ-
omy, including state regulation, is carried out through agents and sectors, that par-
ticipate in the exchange, like commercial banks, trade intermediaries, government
purchases, capital market agents, central banks etc.
Finally we can make the conclusion, that the main difference between the com-
mand and market economy is that the former is outwardly (politically) integrated
while the latter is integrated by different forms of exchange intermediation.
This brief analysis demonstrates, that even if we do not apply any new the-
oretical assumptions and do not use any new information, the application of vi-

---

6
sualization makes generalizations about complex social transformations possible.
These changes a are not so obvious and indeed hard to detect without visualization
techniques.

---