Since the 1990s, financial crises characterized by the currency crisis and banking crisis existing simultaneously occurred in some countries especially developing countries, and this crisis is called twin crisis or double crisis. As the currency crisis theory does not consider the vulnerability of the banking sector, banking crisis theory does not involve opening economic factors, they do not well explain the phenomenon of twin crisis.
Asian financial crisis in 1997 was a typical twin crisis which caused the boom of studying twin crisis in academic circle. Kaminsky and Reinhart (1999) thought that in many countries where currency crisis occurred, their foreign exchange market was attacked, and their domestic banking industry was in crisis. Flood and Marion (2004) proposed a model of the joint distribution of banking and currency crises. They made two points. Firstly, banking and currency crises are related, but they are not the same thing. Viewing crises in isolation or as joint events biases the estimates of the likelihood of crises. Secondly, the proliferation of government promises, such as adding a promise to bail out bank depositors and the promise of fixing the exchange rate, reduces the likelihood of keeping any individual promise when the resources devoted to keeping the promises are scarce. Bleaney et al. (2008) combined a second-generation model of currency crises with a standard banking model. They proposed a model of the interactions between banking crises and currency crises and thought that a run on the banks may cause a currency crisis, or vice versa. There are multiple equilibrium (with either twin crises or no crisis), depending on depositors' expectations of other depositors' actions. Suspension of deposit convertibility can prevent a speculative attack on the currency. Shen (2000) researched the long-term symbiosis and fortuity between the banking and currency crisis by total volume set of method and believed that the banking and currency crisis is a common sport in the long term, the banking crisis would lead to currency crisis, but the currency crisis would not lead to a banking crisis. Liu (2003) analyzed the probability of banking crisis and currency crisis in symbiosis using the frequency distribution and signal method and came to the same conclusion: In emerging market countries, the relationship between the banking crisis and currency crisis does exist, the banking crises even tend to impending currency crisis as a synchronous or early warning indicators, but not vice versa.
Since the majority of academic literatures focuses on the symbiosis of the banking and currency crisis and rarely focuses on how to warn and prevent twin crisis, this paper estimates the probability of twin crisis through empirical analysis. It has certain instructive and realistic significance for constructing perfect twin crisis warning system.
1. MODEL OVERVIEW
This paper estimates the probability of twin crisis through Logit model and Probit model.
1.1 Logit Model and Probit Model Logit model application must meet the following two constraints: The first, assuming a causal relationship between explained variables and explanatory variables; the second, explained variables can only have two choices. The twin crisis this paper considers only has two situations: The occurrence and otherwise. Whether twin crisis occur or not are explained variables, indicated by Y; various factors affecting the occurrence of twin crisis are explanatory variables, indicated by X:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
First considering the linear regression model, the probability of a crisis can be expressed as:
P.sub.i = E (Y = 1|[X.sub.i]) = [[beta].sub.1] + [[beta].sub.2] [X.sub.i])(1)
Where [X.sub.i] is the various influencing factors of twin crisis, and when Y = 1, the twin crisis occur. But now consider the following expression: