In most of the systemic banking crises around the globe that have been scrutinized by economists, East Asia and Latin America crises are always received special attention of scientific world. Not only is this due to it imposes huge fiscal cost to rescue economy, but also their contagious effects exceed any crisis before. One interesting question is that if our model could predict such crises so that contagious effects could be minimized. Theoretically, logit model absolutely fit well such requirements. However, as discussed above, there are always limitations such as data availability or feedback effect of indicators in model building process. Therefore, any conclusion coming from our model is lack of confidence. To access this issue, this chapter begins at chronology description of the 2 cases of East Asia and Latin America crises, Thailand and Uruguay, to clarify real time as well as some important turning points of the crisis, afterwards prediction result will be presented for comparative purpose in next section.
5.1 Thai Banking Crisis:
In the early 1990s, the Bank of Thailand introduced a financial liberalization program, which allowed banks to attract large amounts of foreign capital for Thai’s fast-growing economy. Thai banks and financial companies mobilized short-term debts using banking facilities, called Bangkok International Banking Facilities, and then, converted to bahts at the pegged exchange rate to pour into real estate market. It is believed that the Bank of Thai and the IMF would protect creditors in a crisis, and hence, foreign investors happied to enjoy the higher rates in the Thai local market despite signs that warned the coming crisis. In 1994, the IMF warned Thailand that it needed greater flexibility in its exchange rate regime to slow down the inflow of short-term capital. The central bank, reluctant to put a stop to the impressive economic growth of the preceding years, ignored the warnings. Soon, growth in the real estate sector reached unsustainable levels. Studies from that
- 39 -
period report office vacancy rates in Bangkok exceeding 20 percent in 1996.
There were 300,000 unoccupied new housing units while the annual demand for new housing rarely exceeds 120,000.
In May 1997, the crisis got very close as Thai government effort to save Finance One, the largest Thai finance company, failed. In that period, foreign exchange reserves of central bank were around $38 billion. The most of $30 billion reserves had been committed in forward contracts, and the rest had been used to guarantee the central bank's Financial Institutions Development Fund. In that case, Thai government forced to ignore the government's promise to rescue Finance One with public money, allowing it to bankrupt. This was considered as a renegation of the government’s promise to foreign bankers and depositors that their money would be bailed out. As a consequence, foreign capital flew and generated more pressure for the Thai government to devalue its currency. On 2 July 1997, as speculators convinced that Thai government was out of control the pegged exchange rate, the baths immediately was under attack. Speculative attack led Thailand to abandon its pegged exchange rate regime, following the decline of the Thai stock market and
. many troubles in real estate.
The situation got worse when Thailand government implemented high interest rates and limited fiscal expense to deal with speculative attack on August of 1997. As interest rates started climbing and government spending fell, the economy sank into a deep recession, real estate prices collapsed, and loans made to real estate developers soured. Effects of crisis to Thai economy were severe. According to Caprio and Klingebiel (1996a), nonperforming loans reached the peaked at 33 percent of total loan and were declined to over 10 percent in 2002. Until May 2002, there were more than 50 (of 91) financial companies shut down, account for 13 percent of financial system asset. Fiscal costs estimated at $60 billion, or 42 percent of the GDP in 1999 (Ergungor and Thomson, 2005).
5.2 Uruguay’s crisis.• Victim of contagious effect..•
- 40 -
In 2001, the Uruguay banking system was rated as one of the healthiest ones
in Latin America. The system was probably grouped into two segments, public bank 7 at accounting for 40 percent of the system’s assets, and a group of approximately 30 private mostly foreign banks. Banco Galicia Uruguay (BGU) and Banco Comercial (BC) were two dominated banks in the system, together accounting for approximately 20 percent of the total banking system.
Total Public Banks Private Banks Asset Quality
17.9 39.1 5.6
- NPLs/Total Loans
- Provisions/NPLs 49.7 39.2 91.7
Profitability
-2.3 -4.5 -0.9
- R.O.A. (after tax)
- R.O.E. (after tax) -28.1 35.4 -16.2
Liquidity
93.2 89.5 96.4
- Loans/Deposits
- Liquid Assets/Deposits 15.9 20.9 13.6
Memorandum
100 40.9 59. l
- Total Assets (share)
- FX Deposits/Total Deposits 89.1 84.2 92.1
- FX Loans/Total Loans 80.6 56 93
Soure: Luis and Sophie 2005, page 4.
Figure 1: Breakdown structure of liability side of Uruguay banking system before crisis erupt
US Dollar DcpoxiD by Na’tionality of Depositor
Soure: Luis and Sophie 2005, page 4
- 41 -
7 The public banks include Banco de laRepublica Oriental del Uruguay, or BROU and Banco Hipotecario del Uruguay, or BHU
- 41 -
With liability and assets side highly dollarized (over 80 percent), Uruguayan
banking sector, at that time, became sensitive to external shock and devaluation on
. exchange rate.
The crisis began at the end of 2001 as problems in banking system forced Argentina government to implemented deposit freeze and capital control policies. Immediately, Banco Galicia Uruguay (BGU) and Banco Comercial (BC)- 2 largest private banks - dealt with liquidity problems. During January of 2002, almost BGU liquidity was nearly exshauted with about 15 percent of total deposits lost in this period. In February 2002, the Uruguay central bank decided to shut down BGU pernamently. In fact, what happened to BGU had been predicted. BGU’s liability composition dominated by Argentinean financial groups, thus, it was not surprisingly that as liquidity degree became scarce Argentina investors withdrew their deposits from Uruguay.
Another victim of Argentina crisis that was also worth noting was BC’s case, the largest private bank in Uruguay. Unlike BGU, BC got in trouble in asset side as holding Argentine government debt and lending too much to Argentine borrowers. In first months of 2002, BC also dealt with extensive liquidity tension. After several efforts to restructure, BCU finally interfered BC.
As crisis in Argentine got worse, depositors had moderately left the country.
Until March 2002, deposit withdraws exceeded 12 percent of total bank deposit. Despite liquidity support coming from IMF program, public perspective still believed that currency crisis was going to erupt, which forced the government to widen exchange rate range from 6 to 12 percent. In May 2002, there is a raising concern that Uruguay would follow the footprints of Argentina after its investment status downgraded. Credit market quickly responded to such shock by withdrawing additional 18 percent of deposit. More seriously, the fear spread to local depositors instead of just non-resident ones like before.
To confront these challenges, the central bank continuously provided rescue package to calm down runs on bank. In June 2002, monetary authorities intervened and replaced board of managements of
- 42 -
Bancode Montevideo-Caja Obrera, the third largest private bank in Uruguay.
Following the months of massive withdrawals, foreign exchange reserves were . really low (see the figure below). In December 2001, total reserve declined to 650 million, nearly 80 percent compared to period before the crisis (Luis and Sophie, 2005). Under the pressure of external debt obligation, the Uruguay central bank determined to float its currency and implemented 5-day holidays in July 2002.
Figure 2: Peso/US$ Exchange Rate and Evolution of US$ Reserves (January to December 2002)
Source: Luis and Sophie 2005, page 10
5.3 Implementing the best unbiased model to Thailand and Uruguay cases: failure to explain the crisis originating in balance of payment crisis.
To examine the prediction power of the model, the author uses specification 5a to access crisis probability in Thailand and Uruguay from 1996 to 2002.
The figure below shows that prediction of crisis probability for Thailand is more stable than those of Uruguay. As can be seen, the prediction probability in Thailand reaches the peak of 0.3 in 1997 (the beginning year of Thailand crisis) and decline slowly in following years while the prediction values in Uruguay fluctuate slightly before climb to the top of 0.469 in 2001( the beginning year of Uruguay crisis).
- 43 -
Figure 3: Predicted Probability Of Banking Crisis in Uruguay And Thailand With Specification 5a
Uruguay crisis 2001- 2000
ThaiLand crisis 1997-
0.1
1996 1997 1998 1999 2000 2001 2002 yea rs
Source: Author’s calculation
What we see in the figure tells us that the model obviously makes better prediction on Uruguay crisis than Thailand. If using a fifty percent cut-off(the well- known threshold) to determine whether banking crisis occurs or not, the model completely fails to explain Thai crisis. This result is totally a big surprise. As described in previous section, Thailand economy has begun to show its potential instability in banking system since early 1990s. Under pegged exchange rate regime and belief in protection of IMF and Thai government, Thai market become a fruitful and safe land for international investors to earn profit.
The crisis erupts when speculators realize that the bahts are highly overvalued.
Soon, speculative attacks occur, pegged exchange rate regime collapse, high interest is introduced. One noticeable thing is that the model can not predict Thai crisis in regardless of some macroeconomic conditions during crisis time such as sharp devaluation of exchange rate or growth rate (2 variables highly significant in our model). To address this issue, it is useful to look back marginal effect of exchange rate. In table 3a and 3b, its average marginal effects are quite low, ranging from l .025E-04 to 1.03E-04 for additional unit change of exchange rate. Therefore, prediction power of model is reduced considerably despite speculative attacks on July 1997 led to a surge of exchange rate, moving from 25.61 to 47.65 (nearly 54%ô). Actually, in recent researches about bank crisis such as Demirguc-Kunt and Detragrache (l998a), Hagen and Ho (2002), using absolute value to evaluate effect of exchange rate on
- 57 -
banking sector problem still receives top priority. Although this technique can
" provide empirical evidences for our hypothesis, prediction power of our model is likely to be reduced considerably, especially as handling crises originating from balance of payment crisis.
- 45 -