Dr. BR Ambedkar School of Economics University, Bengaluru VIE Semester § Year Integrated M.Sc. in Economics/ Semester Financial Economics 1d Semester examination~ §.01,2024 Subject N vanced Econometrics Time: 3.00 Hours Max Mari fully, ~ase write down the Section and Serial Number of the question before attempting it. 3. Write legibly and be as concise as possible SECTION ~A Answer any 10 questions each carry 3 marks:(10x3 = 30) e (AR) and Moving Average (MA) models. stationary or not’? . L. Describe the difference between Autoregre: How would you check if a given time serie: . 2. Explain the key difference between Time Series data, cross-sectional data and panel data? tationarity in a time series? Why is it important? Give suitable ex: nples. What is S ~3. Explain seasonality and trend in a time series context. Illustrate with suitable examples the deterministic and nondeterministic time s ies. 4. What is a martingale proces 5. How is ACF and PACF are useful in time series forecasting? 6. “A white noise process is a stationary process. However, all stationary processes are not white noise process.” Discuss. 7. What do you mean by a difference stationary process and trend stationary process? +8. How is the Bounds test useful in ARDL model? 9. Consider a VAR(2) model with two variables, X and Y. Suppose you estimate the model and find that the coefficient on the first lag of Y in the X equation is statistically significant, while the coefficient on the second lag of Y in the X equation is not. However, the coefficient on the first lag of X in the Y equation is not significant, while the coefficient on the second lag of X in the Y equation is significant. What does this pattem of significant coefficients imply about the dynamic relationship between X and Y? 10. You're analyzing two time series, X and Y, which exhibit strong visual similarities: both have upward trends and appear to move together over time. However, when you conduct formal unit root tests (¢.g., Augmented Dickey-Fuller), you find that X is stationary, while Y is non-stationary. Explain how this seemingly contradictory result is possible. What factors could contribute to this discrepancy between visual inspection and unit root tests? 411. How can Fourier analysis and its resulting spectral decompositions shed light on the hidden periodicities and frequency-specific relationships within economic time series data?SE Answer any 5 questions each carry Smarks (5x5= 25) What is the purpose of differencing ARMA and ARIMA. Given below the result table and the figure, interpret the ARCH e! data sch Negra, aren Geeeeing opt tnt zat ton ve. oree Heeration 3 Son: ctere lBeain § Bee Cewttehing opt tatzat beans ar, aad East Sikice: Sem Ronit Sinise: Sest inc tnt rent ete sets oe ee sreus : an eniceeeneacae padecra) > ae SESMUN Gator! mots se Se LS ee a time series dataset? Describe the difference between fects for time series14, <16 17. 19. 20. What are the key advantages of spectral analysis? Write briefly on wavelet analysis and wavelet coherence : Imagine you're analyzing the relationship between coffee consumption (C) K(Q) and sleep duration (S) 1() in a large dataset, Both variables seem intuitively linked: more coffee might lead to shorter sleep, and conversely, sleep deprivation could drive coffee consumption. However, your initial regression analysis reveals that neither series seems stationary on its own (both contain trends). Suggest an econometric technique you could employ to shed light on the association between coffee consumption and sleep duration. Explain the methodology in detail. How would you check any causal relationship? ‘The recent years have witnessed a boom in the tourism industry of Borduria, fueled by its stunning natural beauty and cultural attractions. This surge in tourism coincides with a significant inerease in the import of luxury goods by Borduria. Some economists hypothesize a causal link between these two trends, claiming that the influx of tourists with expensive tastes is driving up the demand for luxury imports. Other economists argue that a common factor, such as a strong economic boom in country A, could be responsible for both the tourism growth and the increased imports. What type of econometric model/models can be used to analyze this phenomenon? Write down a sample model specification as well. Consider a vector autoregressive (VAR) model with two variables, inflation (x) and interest rates (i). Define impulse response functions (IRFs) and variance decomposition (VD) in the context of VAR analysis for these variables. Explain how IRFs and VD can be used to assess the dynamic interactions between inflation and interest rates in this model SECTION -C Answer any 3 questions each carry 10 marks (3x10=30 ) What do you mean by volatility clustering? What is heteroskedasticity in time series analysis? Can you describe the GARCH model and its application in time series analysis? Describe a real-world scenario where a GARCH model can be beneficial. Suppose you're investigating the long-run and short-run relationships between a country's economic growth (Y): 1(0), its investment rate (1): 1(0), and its labor force participation rate (L); I(@) . You've collected time series data on these variables and have decided to employ an ARDL model to analyze their dynamics. Clearly outline the steps involved in constructing an ARDL model in this context, starting from data preparation and ending with model estimation and interpretation. You're an economist examining the dynamic interactions between three key macroeconomic variables in the country Borduria: Real GDP growth (Y), Unemployment rate (U) and Consumer Price Index (CPI) inflation (x). You've collected quarterly data on these variables for the past 20 years and want to model their relationships using a Vectorthe Autoregre VAR) model, You've conducted lag order selection tests and obtained following information criteria values for different lag orders: Lay . . on AIC BIC 1 425.12 | 438.75 ae 418.95 [440.22 y 419.41 | 448.33 [sree 422.15 | 458.72 Hased on the information criteria, which lag order would you choose for your VAR model? Explain your reasoning. Clearly specify the VAR model equation for the chosen lag order, using appropriate notation and indicating the variables and their lags. Provide a brief interpretation of the key elements in the VAR model equation, explaining how it captures the dynamic relationships between the variables, 21. A study analyzed the demand for foreign exchange reserves (res). The study identifies real reserves (Fes), ratio of real imports to GDP (im), ratio of broad money to GDP (m) , exchange rate flexibility (er) and interest rate differential (ird) as factors influencing reserves. The cointegration testing suggests there exists one cointegrating relationship and the study carries out a VECM model and the results are presented below. Explain the VECM method and interpret the following result Variables Dep var: Ares Bem -0.057 (1.98) * | A reset 0.49 (5.9)* | Aim. 1 0.01 (0.10) 0.50 2.979" 0.001 (-0.02) Aird = 0.11 2.32)" Constant 0.025 (2.14)* Re 0.74 | F-stat 19.87 SE. 0} Jarque-Bera 15.47 cates significance at the 5%, Figures in parenthesis and are brackets show t statistics SECTION -DAnswer any 1 question each carry 15 marks ( 1x15=15) >? Ina series, If the variables are independently and identically distributed with a zero mean, have the same variance and co-variance is equal to zero, then how can you you model it? Is this similar or different from “Random Explain the various tests suitable for such interpret stich a series? Walk", explain with suitable examples series. 23.Consider the monthly unemployment rate in a country from January 2010 to December 2023. The time series exhibits seasonality and a slight upward trend. Write the general ARIMA(p, d, q)(P, D, Q) model that could be used to capture these characteristics. Clearly define the parameters and their roles. Describe the steps involved in identifying the appropriate orders (p,d,q) and seasonal orders (P,D.Q) for this model. How would you estimate the ARIMA model? Provide the specific ARIMA model you would estimate, and explain your reasoning for choosing these orders, Write out the mathematical equation forthe ARIMA model if you have chosen ARIMA (1,1, 2) (1, 1, 0). Describe how you would assess the model's goodness-of-fit and forecasting accuracy. Note: Make assumptions wherever required