90
Managing Openness
Table 7.2. Estimation Results, Model of Recession, and GDP Growth in Brazil Recession dummy
Global growth Reserves to external debt Fiscal deficit d (saving ratio) d (terms of trade) Constant Number of observations R-square
GDP growth
Probit
Logit
OLS
–0.82 (0.02)
–1.49 (0.05)
0.48 (0.04)
–0.05 (0.09) 0.38 (0.03) –0.25 (0.03) 0.00 (0.98) 1.63 (0.24)
–0.09 (0.10) 0.64 (0.05) –0.44 (0.05) –0.01 (0.91) 3.03 (0.22)
0.04 (0.03) 0.00 (0.98) 0.40 (0.03) 0.22 (0.00) –0.25 (0.89)
29 0.49
29 0.49
29 0.52
Source: Authors’ calculations. Note: t-statistics are in parentheses. Recession dummy is a variable for Brazilian recessions as indicated by Fundacao Getulio Vargas. OLS = ordinary least squares.
world is in a recession or growing below trend, an increase of one percentage point in global growth reduces by 18 percent the probability of a new Brazilian recession. When the world is growing above trend (higher than 2.5 percent), the impact is much lower: an increase of one percentage point in global growth reduces the probability of a Brazilian recession by just 1.3 percent instead of 18 percent. The third equation is an ordinary least squares version of the first two but with economic growth as the dependent variable. The importance of international reserves in this equation can be explained only if one accepts that instability produced by external shocks has had such an overwhelming effect on growth that its antidote, reserves, seems to be more important at first glance than any other long-run structural elements. In a less shock-prone economy, this result would appear surprising, but in a shattered emerging economy with a long history of macroeconomic instability, it makes perfect sense. External shocks are usually thematic, often related to geographies (Asia, the Russian Federation, Mexico, and so forth) or to specific issues (oil, commodities, or hedge funds in trouble). This time, the theme was broad and far reaching: banks. The worldwide deleveraging and risk-reduction movements produced sudden stops in almost all types of capital inflows, sharp terms-of-trade losses, falling external demand, repatriations of portfolio investments, and even some disruption in foreign trade mechanisms because of the disappearance of trade lines
and sharply increased country risk premium. The direct impact on the exchange rate was brutal: from September onward, in 45 days the real lost approximately 45 percent of its value. Some very unusual influences were behind such a big depreciation of the currency, and the problem was, unfortunately, connected to foreign-exchange derivatives. The surprisingly wide dissemination of “toxic” derivatives products affecting balance sheets of nonfinancial corporations of all sizes multiplied the impacts of the real depreciation on Brazilian companies in novel and dangerous ways. Some interesting parallels are found here with what happened in the United States related to asymmetry of information: some banks sold highly dangerous products to their clients with little explanation of implicit risks, or with outright misrepresentation, and buyers lacked the skills to discuss the products or the capability to manage the risk, or they simply acted on trust. The template is very much the same for the structured products based on subprime mortgages in the United States and for the derivatives based on exchange rates in Brazil. The scant available evidence indicates that a few hundred companies were offered to launch far out-of-themoney put options, which they sold to the banks from which they drew their working capital financing. Thanks to the premium of these options, companies could reduce the costs of bank financing in exchange for the puts they sold, albeit with a very clear underestimation of the risks involved. Notional amounts were huge and potential losses unlimited, if the exchange rate were to depreciate enough to reach the strike prices. If not, it appeared to be only a sophisticated mechanism to reduce the cost of capital. With the unusually large devaluation taking place from September 2008 onward, and at a very rapid pace, the disaster materialized, and these options reached their strike points. Many prime companies entered acute distress, especially in the case of listed companies, where the size of the exposure had to be made public, with truly devastating effects for those involved: Aracruz, Sadia, and Votarantim were the outstanding examples of massive value destruction, while a few hundred other nonpublic companies in the same situation were able to negotiate quietly with their creditors. There was no precise account of how many companies and how much was involved in such operations. Some of the largest banks reported some features of their exposures, with numbers varying from US$500 million to US$1.5 billion in the three banks reporting on these deals. These disclosures unquestionably represented only a fraction of the size of the problem. The Aracruz situation alone (the equity value of the company was around US$9.6 billion in June 2008) involved losses of approximately