THE ACADEMY OF ECONOMIC STUDIES MASTER DAFI

Transcription

1 THE ACADEMY OF ECONOMIC STUDIES MASTER DAFI Econometric modeling of economic crisis influence on mortgage loan Author: Lebada Maria Daniela Coordinator: Conf.dr. Cristian Banacu Prof.univ.dr: Gheorghe Ruxanda Bucharest 2009

2 1. Introduction A major issue for all countries is finding the most adjusted model for financing housing. In the case of Romania, as in other countries, the state has directly funded the real estate sector, where most sources come from savings made by population. Development of the real estate market is closely linked to macroeconomic conditions, to degree of development and efficiency of the financial banking system. In Romania, the main macroeconomic factors are related to lower purchasing power of population and very high level of interest. Influence of global economic crisis on real estate credit market in Romania The people consider that the reason for the current financial crisis is droping of real estate price in U.S. or a collapse of the real estate credit market. This vision is incomplete. Root causes of financial crisis are deep, said many analysts: Altman (2009), Buiter (2008), Blanchard (2009). The main cause of financial crisis was abundant liquidity created by the major central banks of the world (EDF, boj) and the desire of exporting oil and gas countries to limit currency appreciation The consequences of abundant liquidity was very low and volatile interest rates. Both of these consequences have led to increased appetite for assets with large gains. Risk margins was also very low and non-discriminatory. Together, low rates of interest, appetite for assets with high returns, low vigilance against risks and low margins have concealed price signals of the financial markets and led to poor understanding of the risks involved. The crisis. Anti-crisis measures First quarter of 2009 is marked by international financial crisis from the end of last year on global economic activity. 2

3 Interest rates on new loans and deposits rose in January. So, in case of loans, increases were more pronounced, ie 1.43 percentage points (up to a level of percent) in the segment of the retail and 1.68 percentage points (up to a level to percent) on non-financial companies. Interest rates on deposits of individuals have increased to percentage points (+0.79 percentage points), and 0.67 percentage points( percent) on non-financial company. The economic activity worldwide has recorded a substantial loss of rhythm. Demand for exports has declined, and domestic factors, have adversely affected domestic demand. These items correlated with other economic data for the euro area, indicating a negative growth of real GDP in the last quarter of Anti-crisis measures which may be taken in the short term are: measures to counteract the effects of the crisis must be anti-cyclical nature to not escalate the situation created; national anti-crisis measures should be correlated with EU anti-crisis measures level to maximize their positive effects; The process of economic convergence with the EU, including the timetable for accession to the euro area should be maintained; The short term measures should be placed in the medium and long term strategy; economic interests of the population (the purchasing power, capacity for repayment of loans, jobs) and social security support, must be protected; Measures of counteract the effects of the crisis must be taken in a logical order, so that a measure taken to become a resource or support for the nest measures; 2. Case study at a bank Stage of knowledge in econometrics Sir Francis Galton introduced the term of statistical correlation and regression. The econometric modeling procedure is the following : Use of sources: economic theory, experience in modeling, statistical data series; 3

4 Developing the econometric model: specification, identification, measurement, estimation, verification Developing an operational econometric model: results, analysis, forecasting, simulations Confronting reality, updates Presentation of the problem to be solved The main tools used in forecasting management are: forecasting, scheduling, planning and budgeting. Forecasts using scientific models, which mostly used mathematics. But forecasting techniques allow simpler approach, based on observed results, which combine with each other, offering not only a model but rather a "model". Forecast results are used for decisions. For the case study I chose to introduce a unit banking state regarding a particular type of credit, crisis influences on the product, and an analysis of the situation in the banking system regarding the type of mortgage. Construction of the econometric model The general structure of an econometric model is: group of assumptions, economic theory and databases that allow the construction of econometric model a mathematical expression model which usually is an equation or a group of equations containing random multidimensionavariables and / or stochastic processes. Multifactorial model is defined by the following relationship y = f (x j ) + u where: y = dependent variable x j = independent variable j = 1, k, k = number of exogenous variables; 4

5 u = random variable The mathematical expression of the econometric model is the regression function: Y= a 0 +a 1 X 1 +a 2 X 2 +ξ The result of the analysis of the relation between interest rate and the exchange rate is a coefficient of determination of This means that the interest rate is explained by the exchange rate in proportion of 85%. A coefficient of determination of 0.54 was calculated to estimate the relation between the interest rate and the inflation rate That means that interest rate is explained by inflation rate in proportion of 54%, less than exchange rate. The result of the relation between the interest rate and two variables: the exchange rate and the inflation rate, is given by a coefficient of determination of 0.88, meaning that these rates explain the exchange rate evolution in proportion of 88%. In order to make a correct analysis, Multicollinearity tests must be applied. Multicollinearity occurs when two (or more) variables are related 8 CURS_VAL vs. RATA_INF RATA_INF = * CURS_VAL Correlation: r = RATA_INF CURS_VAL Regression 95% confid. 5

6 Following the application of the Klein, Farrer and Glauber tests, the existence of multicollinearity is confirmed, which is corrected by keeping the model with a greater coefficient of determination The Durbin Watson test indicated the presence of positive autocorelation in residuals from the regression analysis. 24 Distribution of Raw residuals No of obs Expected Normal Forecasts using ARIMA model (autoregresiv mobile average) The ARIMA model for forecasting was used as a method for the detection of stationary or nonstationary series. The graph below shows a downward trend, not a constant value, so the series is nonstationary. 6

7 80 Plot of variable: RATA_DOB RATA_DOB Case Numbers Will be serious differences in the first step, then I ll present autocorelaţions function and corelogram errors (standard derrors are white-noise estimates) Lag Corr. S.E Autocorrelation Function RATA_DOB Q p

8 Lag Corr. S.E Autocorrelation Function RATA_DOB: ARIMA (1,2,0) residuals; (Standard errors are white-noise estimates) Q p Evolution of the mortgage at banks from system The purpose of this analysis is to find the relevant indicators for a more accurate ranking of the banking units For this purpose, the analysis used a correlation Matrix to see if the calculated indicators are independent or not. Correlation Matrix describes the links between initial variables and also determine the number of principal components. 8

9 Correlation Matrix(a) Portofoliul Zscore rata cost de credite (Active) suma max per max rata dob DAE estim lunara total Portofoliul de credite 1,000,142,198,235,067,000,144,144 Zscore(Active),142 1,000,157,213,041,120,017,016 suma max,198,157 1,000,051,367,385,266,291 per max,235,213,051 1,000,002,171,265,249 rata dob,067,041,367,002 1,000,777,533,544 DAE estim,000,120,385,171,777 1,000,595,571 rata lunara,144,017,266,265,533,595 1,000,993 cost total,144,016,291,249,544,571,993 1,000 a Determinant =,001 Other methods of determining the number of principal components are table Total variance Explained and Screen Plot-ul 4. Results Interpretation The correlation between initial variables and the last 3 components obtained by the methods described above is important for for the principal components analysis. This is indicated in Component Matrix For a better interpretation was used Rotated Component Matrix. Following this analysis, the conclusion is that the main factors to be considered in choosing a unit bank are: interest rate, DAE, monthly rate and total cost of loan. 9

10 Rotated Component Matrix(a) Component Portofoliul de credite,001,775,404 Zscore(Active),081,022,804 suma max,462,017,545 per max,153,728,436 rata dob,833,139,161 DAE estim,877,074,047 rata lunara,855,302,046 cost total,852,296,008 Extraction Method: Principal Component Analysis. Component Score Covariance Matrix, which show the links between important components Can be notice that, regardless of method used, banking establishments are hierarchical in 2 groups, if we consider the maximum level of aggregation. Component Score Covariance Matrix Component ,000,000,000 2,000 1,000,000 3,000,000 1,000 10

11 5. Conclusions and recommendations As illustrated by the graphs presented, the terms of mortgage lending have been modified in the last quarter of 2008, with regard to: maximum weight of debt service in monthly income the spread of the average rate of interest compared to ROBOR 1M the maximum share of credit in the amount of the real-estate warranty In order to prevent bank risk, banks must comply with several regulations. In general, these regulations on banking activities have a message of protection against multilateral risk resonance. For example, are protecting: - Banks from the risks caused by quality clients and their hostile actions; - Clients against incompetent bank management; - The interests of society against instability arising from unbalance banks' management with a large appetite for risk. Efforts should be intensified during this period, mainly to convince customers that the bank will provide them with stability, qualified staff and suitable products for their specific needs For this reason, in promoting their services, banks should act by the motto "Promise only what you can deliver and then deliver more than you promised Bibliography: 1. Banacu Cristian, Aplicatii in evaluarea investitiilor cu ajutorul tabelelor de calcul actuarial, Tribuna Economica, Bucuresti, Banacu Cristian, Analiza si evaluarea afacerii, Tribuna Economica, Bucuresti, Brasoveanu Laura, Brasoveanu I, Corelatia dintre politica fiscala si cresterea economica, Economie teoretica si aplicata v. 15, nr. 7, Ciobanu Ana Maria, Stancu I, Analiza performantelor intreprinderii, Editura ASE, Bucuresti,

12 5. Dardac N, Basno C Monedă, credit, bănci, Editura Didactică şi pedagogică, Bucureşti, Dardac, N., Basno C, Riscurile bancare : Cerinţe prudenţiale. Monitorizare 7. Dobrotă N, Economie Politică, Editura Economică, Bucureşti, Ruxanda G., Note de curs 2009, Editura ASE, Bucuresti, Ruxanda G., Analiza datelor, Editura ASE, Bucureşti, Stancu I, Piete financiare si gestiunea portofoliului, Editura Economica,Bucuresti, Stancu S, Modelarea matematica a sistemelor si proceselor de conducere : analiza sistemelor economice complexe : teorie si aplicatii, Editura ASE, Bucuresti, Stancu S, Andrei T, Statistică Teorie şi aplicaţii, Editura All, Bucureşti, Spircu L., (2005), Analiza datelor: aplicatii economice, Editura ASE, Bucuresti. 14. Trebici V, Mica enciclopiedie de statistică, Editura Ştiinţifică şi pedagogică, Bucureşti,2005 xxx