### Table 5: Explained and unexplained variances in the fluctuations

"... In PAGE 14: ... The models in Table 4 explain over 90% of the variance across states and over time in the poverty measures. However, a large share of this explained variance is attributable to the state- specific intercepts and time trends ( Table5 ). It is more difficult to explain the fluctuations.... ..."

### Table 4 Akaike Information Criterion Values for Markov Switching Models 2 regimes 3 regimes 4 regimes

"... In PAGE 12: ... 12 The Akaike Information Criterion (AIC) values for 2 to 4 regimes and 1 to 4 lags are shown in Table4 . The results show that the lowest AIC value corresponds to the Markov regime switching model with 3 regimes and 1 lag.... ..."

### Table 4: Determinants of the fluctuations in rural poverty measures

"... In PAGE 13: ... 3 We use a nonlinear least squares estimator of model (1)-(2). 4 The estimated parameters for the key time-dependent variables are reported in Table4 for two versions of the model, with and without current development spending.... In PAGE 14: ... The current level of development spending itself turned out to be insignificant in all equations. The models in Table4 explain over 90% of the variance across states and over time in the poverty measures. However, a large share of this explained variance is attributable to the state- specific intercepts and time trends (Table 5).... In PAGE 17: ... These extensions included: (i) introducing current real development spending as an additional regressor in the model, (ii) allowing for a nonlinear (quadratic) state-specific time-trend, (iii) including lagged real agricultural wage as an additional explanatory variable, (iv) allowing a quadratic term in the rate of inflation, (v) allowing for state-specific effects of inflation and the real wage rate. The parameter estimates for the model with current development spending are given in Table4 , which shows the current levels of development spending to be insignificant. The inclusion of this variable did not improve the predictions for 1992 either.... ..."

### TABLE 1 ILLUSTRATIVE VALUES FOR 3 FLUCTUATIONS IN THE SCDM MODEL

### Table 13 Regime-switching Model (Thailand)

"... In PAGE 29: ... The model is estimated using four time series: the index returns of a crisis source country, the investable and non-investable index returns of another country, and the world index. We report the model results in Table13 with Thailand as the crisis source country. For each country in the table we separately estimate the parameters of the regime switching model outlined in Section 3.... In PAGE 29: ...28 Changes in volatility have been annualized. Results for emerging market countries are given in panel A of Table13 while results for developed countries are given in panel B. The columns to the left (1a-10a) of Table 13 report differences in estimated moments while the columns to the right (1b-10b) report the corresponding t-statistics to test the restrictions defined by Equations 13 to 15.... In PAGE 29: ...o not report levels.28 Changes in volatility have been annualized. Results for emerging market countries are given in panel A of Table 13 while results for developed countries are given in panel B. The columns to the left (1a-10a) of Table13 report differences in estimated moments while the columns to the right (1b-10b) report the corresponding t-statistics to test the restrictions defined by Equations 13 to 15. The first 4 columns, columns 1a through 4a, report differences in volatility to test restriction 13, columns 5a through 8a report differences in correlations to test restriction 14, and columns 9 and 27Interestingly, when we separate the investable returns from the non-investable returns (Table 12B), the increasing dependence pattern disappears for Czech Republic, Greece, and Hungary.... In PAGE 30: ... Hence, the country crisis regime for Turkey appears to be a period of isolated country-specific economic turmoil that occurs over periods when markets are relatively calm in the rest of the world. However, for a majority of the coun- tries in Table13 , the crisis regimes appear to be associated with economic turmoil in Thailand. Of the 44 countries, 36 have a positive significant dif- ference in volatility for the investable return and Thailand (columns 3a and 1a, respectively).... In PAGE 64: ...000 0.000 Correlation Investable Investable Volatility Correlation Volatility Table13 . Continued Difference In Estimated Moments Across States... ..."

### Table 10 Identifying Regimes by Model Free Characteristics (Basis Points)

"... In PAGE 28: ... In general, considerable caution should be exercised in interpreting the regime classification as there is considerable error in this classification due to parameter uncertainty and model misspecification.16 Table10 provides some simple statistics relating to the 6-month and 5-year yields, ob- tained from the simulated yields from the preferred 2-Factor[RS] model. The different sta- tistics are computed by regime s = 0, 1.... ..."

### Table 1: Probability of regime transitions.

"... In PAGE 14: ... This allows us to report actual transition probabilities Pij as opposed to the previous reports of transition counts Tij only. Table1 shows the estimated transition probability for all the regime episodes regardless of regime duration, as well as for regime events whose duration is of 6 days or longer. For both cases, the transition probability from the low- or high-latitude regime back to the regime itself is highest, while the probabilities of transition from Regime 1 (central-peak regime) to all three regimes (including itself) are fairly similar.... In PAGE 15: ... It thus appears that Regimes 2 and 3, while associated with opposite polarities of EOF-1, are not just opposite phases of an oscillation in the jet apos;s latitudinal position.[ Table1 near here, please ] Although the regime composites in Fig. 8 might suggest a single zonal jet shifting merid- ionally, this impression is rather super cial and misleading.... In PAGE 15: ... As evident from Fig. 9, the zonal- ow vacillation in our model is not characterized simply by the meridional displacement of a single jet, but by more complex changes in the meridional wind pro le that are often associated (see also Table1 ) with transitions between single- and double-jet states. The regimes apos; persistence characteristics are summarized in Fig.... In PAGE 26: ...haracterized by a strong single jet with enhanced meridional shear (Figs. 8 and 9). This asymmetry in jet structure suggests that zonal- ow vacillation does not simply involve merid- ional shifts in the jet apos;s position and changes in its intensity. Given the length of our model simulations, we were able to compute transition probabil- ities between regimes ( Table1 ). The transition matrix supports the idea|already emerging from the asymmetry of the jet structure in the two o -climatology regimes|that the vacil- lation is not a manifestation of linear oscillations about the climatological mean jet.... ..."

### Table 2.1 PDF and CHF of fading amplitudes for several fading channel models

2003

### Table 1: FSC Models Representing Lognormal Fading Channels

1996

"... In PAGE 3: ... The state probabilities of the FSC were so chosen that the average BER for each state of the model matches that of the corresponding range of the fade depth of the original channel. The resulting FSC models are shown in Table1 . Here, we are not concerned with the state transition probabilities since we have assumed su ciently slow fading.... In PAGE 3: ... In all cases investigated, the model selects the optimal (rs; rc) pair. Table1 shows the di erent FSC models representing lognormal fading channels. The FSC apos;s for the Rayleigh channel were very similar to those for the lognormal chan- nels and therefore we chose to present results for the lat- ter case only.... In PAGE 3: ... This is also the most relevant case for the problem at hand since Rayleigh fading is usually faster than lognormal and does not lend itself easily to adaptive transmission. For each case in Table1 , we investigate the performance of the system for memoryless Gaussian sources as well as Gauss-Markov sources with correlation coe cient = 0:9. We consider di erent dimensions for the TSVQ and di erent transmission rates for the chan- nel as summarized in Tables 2-5.... ..."

Cited by 15

### Table 1: FSC Models Representing Lognormal Fading Channels

1996

"... In PAGE 3: ... The state probabilities of the FSC were so chosen that the average BER for each state of the model matches that of the corresponding range of the fade depth of the original channel. The resulting FSC models are shown in Table1 . Here, we are not concerned with the state transition probabilities since we have assumed su ciently slow fading.... In PAGE 3: ... In all cases investigated, the model selects the optimal (rs; rc) pair. Table1 shows the di erent FSC models representing lognormal fading channels. The FSC apos;s for the Rayleigh channel were very similar to those for the lognormal chan- nels and therefore we chose to present results for the lat- ter case only.... In PAGE 3: ... This is also the most relevant case for the problem at hand since Rayleigh fading is usually faster than lognormal and does not lend itself easily to adaptive transmission. For each case in Table1 , we investigate the performance of the system for memoryless Gaussian sources as well as Gauss-Markov sources with correlation coe cient = 0:9. We consider di erent dimensions for the TSVQ and di erent transmission rates for the chan- nel as summarized in Tables 2-5.... ..."

Cited by 15