ESCALA DE ACTITUDES DE LOS PROFESORES HACIA LA INCLUSIÓN
DATOS DEL PROFESOR (*)
In the first stage, default or restructuring is equivalent to denying obligations, prepayment in cash or defeasance represent prepayment options, and partial pre- payment includes curtailment or partial defeasance. We set up four models where we: (1) exclude property demand factors, mortgage market rates and Treasury yield; (2) include property demand factors only; (3) include mortgage market rates and Treasury yield; (4) include all property demand factors, mortgage mar- ket rates and Treasury yield.
Tables 5.7 and 5.8 report the main results. Findings are consistent with theoret- ical expectations except for the effect of credit tenant lease on the likelihood of denying obligations and the impact of occupancy rate on the likelihood of partial prepayment. Based on the Bayesian Information Criteria, Model 4 is the best fit. Among mortgage terms, negative amortization rate, mortgage rates and mort- gage type (interest only), exhibit statistically and economically significant impact on the likelihood of obligation denial: a 1% increase in each of these variables determines respectively a 91% decrease and 110% and 63% increase in the prob- ability. Despite cross-collateralization being a protection covenant for lenders, we find that this phenomenon does not reduce the likelihood of obligation denial while, instead, a 1% increase in cross-collateralization makes the probability in- crease by 18%. This result reflects the endogeneity of the underwriting process. Furthermore, prepayment penalty / yield maintenance works as a tool to prevent prepayment as it causes a 2% reduction in the likelihood of obligation denial. Importantly, the LTV ratio shows a significant impact on the decision to deny obligations, with a 1% change causing a 34% shift in the probability of default. Hence, the careful consideration banking regulators assign to this metric is justi- fied, especially considering the impact on banking losses in case of option exercise. Furthermore, other collateral related terms also exert a significant impact on the decision of denying obligations: a 1% increase in DSCR reduces the likelihood of
denying obligation by 21%.
Finally, we find evidence of our main hypothesis that property supply constraints is important and consequently banking regulators should carefully consider the geographical composition of the lending portfolio. If the supply elasticity of a market increases by one unit, the likelihood of denying obligation drops by 55%. Therefore, relaxing supply constraints can reduce the possibility of denying obli- gation and can be seen as an alternative risk management tool to control systemic default risk and spillover effects.
Table 5.7: Two-Stage Model (Winsorized): The First Stage -Multinomial Logit Model Results - Coefficients Independent Variables M1 M2 M3 M4 D or R P or Df C or PD D or R P or Df C or PD D or R P or Df C or PD D or R P or Df C or PD LT Vm,t 0.127*** 0.007*** 0.010*** 0.127*** 0.007*** 0.010*** 0.289*** 0.010*** 0.021*** 0.289*** 0.010*** 0.021*** (0.001) (0.0004) (0.003) (0.001) (0.0004) (0.003) (0.002) (0.0004) (0.003) (0.002) (0.0004) (0.003) T T Mm,t -0.005*** -0.005*** 0.010*** -0.005*** -0.005*** 0.010*** -0.004*** -0.005*** 0.010*** -0.004*** -0.005*** 0.010*** (0.0001) (0.0001) (0.0002) (0.0001) (0.0001) (0.0002) (0.0001) (0.0001) (0.0002) (0.0001) (0.0001) (0.0002) BP Rm,t 0.005*** -0.018*** -0.002** 0.005*** -0.018*** -0.002** 0.007*** -0.018*** -0.001 0.007*** -0.018*** -0.001 (0.0002) (0.0001) (0.001) (0.0002) (0.0001) (0.001) (0.0003) (0.0001) (0.0008) (0.0003) (0.0001) (0.001) N ARm,t -2.069*** 0.091 -2.260*** -2.061*** 0.092 -2.252*** -2.251*** 0.080 -2.491*** -2.252*** 0.080 -2.492*** (0.253) (0.079) (0.259) (0.253) (0.079) (0.259) (0.287) (0.081) (0.291) (0.287) (0.081) (0.291) M Rm,t 0.478*** 0.474*** 0.410*** 0.477*** 0.474*** 0.410*** 0.741*** 0.529*** 0.591*** 0.741*** 0.530*** 0.591*** (0.009) (0.004) (0.033) (0.009) (0.004) (0.033) (0.010) (0.004) (0.033) (0.010) (0.004) (0.033) M Tm 0.609*** -0.237*** 0.380*** 0.609*** -0.237*** 0.380*** 0.490*** -0.223*** 0.402*** 0.490*** -0.223*** 0.402*** (0.008) (0.007) (0.033) (0.008) (0.007) (0.033) (0.009) (0.007) (0.034) (0.009) (0.007) (0.034) CCm 0.191*** 0.163*** -0.879*** 0.191*** 0.163*** -0.879*** 0.166*** 0.169*** -0.871*** 0.166*** 0.170*** -0.871*** (0.028) (0.020) (0.219) (0.028) (0.020) (0.219) (0.031) (0.020) (0.222) (0.031) (0.020) (0.222) P P Y Mm,t -0.030 -215.5425 0.040 -0.030 -203.265 0.040 -0.022 -340.769 0.043 -0.022 -340.807 0.043 (0.034) (3920.91) (0.045) (0.034) (3319.07) (0.045) (0.036) (22545) (0.046) (0.036) (22544) (0.046) DSCRm,t -0.589*** -0.282*** 0.528*** -0.585*** -0.282*** 0.527*** -0.240*** -0.232*** 0.800*** -0.240*** -0.231*** 0.800*** (0.017) (0.010) (0.062) (0.017) (0.010) (0.062) (0.019) (0.010) (0.062) (0.019) (0.010) (0.062) P Am,t -0.002*** 0.007*** -0.003*** -0.002*** 0.007*** -0.003*** -0.0001 0.007*** -0.004*** -0.0001 0.007*** -0.004*** (0.0003) (0.0002) (0.001) (0.0003) (0.0002) (0.001) (0.0003) (0.0002) (0.001) (0.0003) (0.0002) (0.001) ORm,t -0.033*** 0.018*** 0.007*** -0.033*** 0.018*** 0.007*** -0.037*** 0.013*** -0.004** -0.037*** 0.013*** -0.004** (0.0003) (0.0004) (0.002) (0.0003) (0.0004) (0.002) (0.0004) (0.0004) (0.002) (0.0004) (0.0004) (0.002) CT Lm -0.326*** -1.409*** -19.742 -0.326*** -1.409*** -19.375 0.031 -1.393*** -26.267 0.031 -1.393*** -26.268 (0.118) (0.098) (4204.59) (0.118) (0.098) (3496.88) (0.134) (0.099) (113345) (0.134) (0.099) (113357) P SEm -0.999*** 1.487*** -0.908*** -0.982*** 1.491*** -0.907** -0.814*** 1.519*** -0.647 -0.800*** 1.523*** -0.647 (0.088) (0.065) (0.461) (0.089) (0.066) (0.461) (0.098) (0.066) (0.473) (0.098) (0.066) (0.473) ∆P ARm,t−3 -0.001*** -0.0007** 0.0004 -0.0007** -0.0006** -0.00002 (0.0003) (0.0003) (0.003) (0.0983) (0.0003) (0.003) ∆F M Rt−3 3.989*** 3.942*** 3.989*** 3.990*** 3.943*** 3.990*** (0.302) (0.191) (0.827) (0.302) (0.191) (0.827) ∆IT Yt−3 -0.078*** -0.076*** -0.071*** -0.078*** -0.076*** -0.071*** (0.008) (0.005) (0.021) (0.008) (0.005) (0.021) Constant -8.669 4.454 -12.046 -8.667 4.451 -12.049 -13.670*** 5.828*** -0.161 -13.660*** 5.827*** -0.162 (368.935) (58.346) (2092.91) (368.192) (58.336) (2090.51) (0.898) (0.577) (2.456) (0.898) (0.577) (2.456)
Fixed State Effects Y Y Y Y Y Y Y Y Y Y Y Y
Fixed Time Effects Y Y Y Y Y Y Y Y Y Y Y Y
BIC 754364 754379 725464 725341
Log Likelihood -372619 -372606 -358350 -358268
LR Chi Square 276988 276717 287774 287649
Observations 989851 989471 964144 963766
Notes:
LT Vm,t: LTV ratio,T T Mm,t: Time to Maturity,BP Rm,t: Balloon Payment Ratio,N ARm,t: Negative Amortization Rate,M Rm,t: Mortgage Rate,M Tm: Mortgage Type (Interest Only),
CCm: Cross Collateralization,P P Y Mm,t: Prepayment Penalty & Yield Maintenance,DSCRm,t: Debt Service Coverage Ratio,P Am,t: Property Age,ORm,t: Occupancy Rate,CT Lm:
Credit Tenant Lease,P SEm: Property Supply Elasticity,OEm,t: Office Using Employment in MSA,P ARm,t: Property Absorption Rate in MSA,F M Rt: Fixed Rate Mortgage Market Rate. D, R, P, Df, C and PD mean Default, Restructured, Prepaid, Defeased, Curtailed and Partially Defeased.
Figures in parentheses are standard errors.
***, ** and * represent significant level at 1%, 5%, 10%.
Table 5.8: Two-Stage Model (Winsorized): The First Stage - Multinomial Logit Model Results - Relative Risk Ratio Independent Variables M1 M2 M3 M4 D or R P or Df C or PD D or R P or Df C or PD D or R P or Df C or PD D or R P or Df C or PD LT Vm,t 1.136*** 1.007*** 1.010*** 1.136*** 1.007*** 1.010*** 1.335*** 1.010*** 1.022*** 1.336*** 1.010*** 1.022*** T T Mm,t 0.995*** 0.995*** 1.010*** 0.995*** 0.995*** 1.010*** 0.996*** 0.995*** 1.010*** 0.996*** 0.995*** 1.010*** BP Rm,t 1.005*** 0.982*** 0.998** 1.005*** 0.982*** 0.998** 1.007*** 0.982*** 0.999 1.007*** 0.982*** 0.999 N ARm,t 0.126*** 1.096 0.104*** 0.127*** 1.096 0.105*** 0.105*** 1.083 0.083*** 0.105*** 1.083 0.083*** M Rm,t 1.612*** 1.606*** 1.507*** 1.612*** 1.606*** 1.507*** 2.098*** 1.698*** 1.805*** 2.099*** 1.698*** 1.806*** M Tm 1.838*** 0.789*** 1.463*** 1.838*** 0.789*** 1.462*** 1.632*** 0.800*** 1.495*** 1.632*** 0.800*** 1.495*** CCm 1.210*** 1.176*** 0.415*** 1.211*** 1.177*** 0.415*** 1.181*** 1.185*** 0.419*** 1.181*** 1.185*** 0.419*** P P Y Mm,t 0.970 0 1.041 0.970 0 1.041 0.978 0 1.044 0.978 0 1.044 DSCRm,t 0.555*** 0.754*** 1.695*** 0.557*** 0.755*** 1.694*** 0.786*** 0.793*** 2.225*** 0.786*** 0.793*** 2.226*** P Am,t 0.998*** 1.007*** 0.997*** 0.998*** 1.007*** 0.997*** 1.000 1.007*** 0.996*** 1.000 1.007*** 0.996*** ORm,t 0.967*** 1.018*** 1.007*** 0.967*** 1.018*** 1.007*** 0.964*** 1.013*** 0.996** 0.964*** 1.013*** 0.996** CT Lm 0.722*** 0.244*** 0 0.722*** 0.244*** 0 1.031 0.248*** 0 1.032 0.248*** 0 P SEm 0.368*** 4.422*** 0.403*** 0.375*** 4.441*** 0.404** 0.443*** 4.568*** 0.524 0.449*** 4.585*** 0.524 ∆P ARm,t−3 0.999*** 0.999** 1.0004 0.999** 0.999** 1.000 ∆F M Rt−3 54.001*** 51.542*** 54.023*** 54.079*** 51.566*** 54.038*** ∆IT Yt−3 0.925*** 0.927*** 0.932*** 0.925*** 0.927*** 0.932*** Constant 0.0002 85.963 5.87×10−6 0.0002 85.720 5.86×10−6 1.16×10−6*** 339.648*** 0.851 1.17×10−6*** 339.242*** 0.850
Fixed State Effects Y Y Y Y Y Y Y Y Y Y Y Y
Fixed Time Effects Y Y Y Y Y Y Y Y Y Y Y Y
Notes:
LT Vm,t: LTV ratio,T T Mm,t: Time to Maturity,BP Rm,t: Balloon Payment Ratio,N ARm,t: Negative Amortization Rate,M Rm,t: Mortgage Rate,M Tm: Mortgage Type (Interest Only),
CCm: Cross Collateralization,P P Y Mm,t: Prepayment Penalty & Yield Maintenance,DSCRm,t: Debt Service Coverage Ratio,P Am,t: Property Age,ORm,t: Occupancy Rate,CT Lm: Credit Tenant Lease,P SEm: Property Supply Elasticity,OEm,t: Office Using Employment in MSA,P ARm,t: Property Absorption Rate in MSA,F M Rt: Fixed Rate Mortgage Market Rate.
D, R, P, Df, C and PD mean Default, Restructured, Prepaid, Defeased, Curtailed and Partially Defeased. ***, ** and * represent significant level at 1%, 5%, 10%.
At the same time, significant impacts of mortgage terms are obtained in prepay- ment options. The likelihood of prepayments respectively increases by 71% when rates increase by 1% and decreases by 20% for interest-only mortgages. Further- more, as zero prepayment penalty and yield maintenance are found in prepaid and defeased mortgages (refer to table 5.3), we obtain some “zero” relative risk ratios. This finding is consistent with market practice, where prepayment in cash normally happens during the last few months of the contract when the bank does not ask for penalties and normally renegotiates the contract if the property is not sold. Similar to obligation denial, collateral related terms such as DSCR and credit tenant lease generate significant impact on the prepayment probability (re- duction of 21% and 75% for respectively a 1% increase in DSCR and when an agreement is signed.
Finally, if property supply elasticity rises up by one unit, the prepayment prob- ability will be raised to more than three times. This reflects that borrowers in elastic supply area have stronger willingness to fully prepay in order to maximise their profits because they can predict to a less degree of uncertainty than in sup- ply constrained markets.
Tables 5.9 and 5.10 report the estimation results for the second stage: default vs restructuring, prepayment in cash vs defeasance, and curtailment vs partial defeasance. Model M4a shows how a borrower’s choice between the two denial obligation options are affected by mortgage and collateral related terms. LTV ratios, time to maturity and balloon payment rate have minimal impact, while a 1% change in negative amortization rate and mortgage rate significantly affect the likelihood of restructuring vs default (respectively by 49% and 42%). Importantly, interest only mortgages show a probability which is three times bigger due to the high level of outstanding loan balance present at the time of exercising the option in comparison with constant payment mortgages. Moreover, if a mortgage contract includes a cross-collateralization covenant, the likelihood of restructuring relative to default is 23% higher. As far as collateral related terms are concerned, increasing the DSCR by 1% increases the likelihood of restructuring by 47%. Furthermore, property supply constraints also exert a significant impact and a 73% reduction in the probability of restructuring is obtained when the supply elasticity increases by one unit. As the restructuring becomes less likely to occur for a mortgage collateralized with an office building in supply elastic markets, we infer that lenders probably expect a greater bounce of property prices in supply inelastic markets and they are therefore willing to modify the loan structure for temporary insolvent borrowers more than for collateral assets in supply elastic
Table 5.9: Two-Stage Model (Winsorized): The Second Stage - Logit Model Results - Coefficients
Independent Variables M1a M1b M1c M2a M2b M2c M3a M3b M3c M4a M4b M4c
LT Vm,t -0.079*** 0.003 0.128*** -0.079*** 0.003 0.128*** -0.048*** 0.003* 0.127*** -0.072*** 0.003* 0.128*** (0.002) (0.002) (0.021) (0.002) (0.002) (0.021) (0.002) (0.002) (0.021) (0.002) (0.002) (0.021) T T Mm,t 0.0002 -0.0007*** -0.002 0.0002 -0.0007*** -0.002 −9.48×10−6 -0.0007*** -0.002 0.0001 -0.0006*** -0.002 (0.0003) (0.0002) (0.002) (0.0003) (0.0002) (0.002) (0.0003) (0.0002) (0.002) (0.0003) (0.0002) (0.002) BP Rm,t -0.004*** -0.017*** -0.004*** -0.017*** -0.004*** -0.017*** -0.004*** -0.017*** (0.0006) (0.005) (0.0006) (0.005) (0.0006) (0.005) (0.0006) (0.005) N ARm,t -0.682*** 21.785* -0.682*** 21.908* -0.642*** 21.565* -0.670*** 21.704* (0.092) (12.636) (0.092) (12.704) (0.092) (12.620) (0.092) (12.686) M Rm,t -0.610*** 0.225*** 0.610*** -0.610*** 0.224*** 0.605*** -0.359*** 0.228*** 0.608*** -0.539*** 0.227*** 0.603*** (0.024) (0.015) (0.171) (0.024) (0.015) (0.172) (0.022) (0.015) (0.171) (0.024) (0.016) (0.171) M Tm 1.096*** -0.625*** 0.894*** 1.097*** -0.625*** 0.898*** 1.095*** -0.627*** 0.879*** 1.099*** -0.626*** 0.884*** (0.020) (0.034) (0.208) (0.020) (0.034) (0.209) (0.020) (0.034) (0.208) (0.020) (0.034) (0.209) CCm 0.201*** 0.430*** 0.2003*** 0.431*** 0.198*** 0.430*** 0.205*** 0.430*** (0.060) (0.062) (0.060) (0.062) (0.060) (0.062) (0.060) (0.062) P P Y Mm,t -0.002 -0.001 0.009 0.001 (0.082) (0.081) (0.076) (0.080) DSCRm,t 0.343*** 0.111** -1.500*** 0.342*** 0.111** -1.501*** 0.506*** 0.119** -1.470*** 0.383*** 0.118*** -1.471*** (0.035) (0.052) (0.297) (0.035) (0.052) (0.296) (0.034) (0.052) (0.297) (0.035) (0.052) (0.296) P Am,t 0.005*** 0.006*** 0.021*** 0.005*** 0.006*** 0.021*** 0.005*** 0.006*** 0.021*** 0.005*** 0.006*** 0.021*** (0.0007) (0.0006) (0.004) (0.0007) (0.0006) (0.004) (0.0007) (0.0006) (0.005) (0.0007) (0.0006) (0.004) ORm,t 0.015*** 0.016*** -0.056*** 0.015*** 0.016*** -0.056*** 0.016*** 0.015*** -0.055*** 0.016*** 0.015*** -0.055*** (0.0008) (0.001) (0.009) (0.0008) (0.001) (0.009) (0.0008) (0.001) (0.009) (0.0008) (0.001) (0.009) CT Lm -1.229 3.801*** -1.225 3.798*** -1.135 3.801*** -1.025 3.798*** (1.024) (1.320) (1.024) (1.319) (1.024) (1.318) (1.026) (1.317) P SEm -1.359*** 1.837*** -13.273*** -1.355*** 1.844*** -13.498*** -1.151*** 1.827*** -13.071*** -1.291*** 1.834*** -13.292*** (0.210) (0.237) (1.906) (0.210) (0.237) (1.925) (0.209) (0.238) (1.897) (0.212) (0.238) (1.916) ∆P ARm,t−3 -0.001 -0.001 0.023 -0.001 -0.001 0.023 (0.001) (0.001) (0.017) (0.001) (0.001) (0.017) ∆F M Rt−3 -0.008 1.167 5.779** -0.072 1.166 5.787** (0.131) (0.872) (2.925) (0.123) (0.872) (2.932) ∆IT Yt−3 0.001 -0.023 -0.119 0.006 -0.023 -0.119 (0.014) (0.023) (0.077) (0.013) (0.023) (0.077) Constant 6.449*** -6.224*** -5.286** 6.444*** -6.219*** -5.170** 0.606 -3.019 10.461 4.047*** -3.016 10.603 (0.983) (0.465) (2.526) (0.983) (0.465) (2.528) (0.384) (2.622) (8.987) (0.374) (2.622) (9.005)
Fixed State Effects Y Y Y Y Y Y Y Y Y Y Y Y
Fixed Time Effects Y Y Y Y Y Y Y Y Y Y Y Y
Log Likelihood -20146 -16234 -306 -20140 -16227 -305 -20016 -16132 -304 -19825 -16126 -303
LR Chi Square 14667 11813 467 14665 11802 468 14036 11674 456 14406 11663 457
Observations 47512 36739 1652 47500 36728 1652 46731 36382 1626 46719 36371 1626
Notes:
LT Vm,t: LTV ratio,T T Mm,t: Time to Maturity,BP Rm,t: Balloon Payment Ratio,N ARm,t: Negative Amortization Rate,M Rm,t: Mortgage Rate,M Tm: Mortgage Type (Interest Only),
CCm: Cross Collateralization,P P Y Mm,t: Prepayment Penalty & Yield Maintenance,DSCRm,t: Debt Service Coverage Ratio,P Am,t: Property Age,ORm,t: Occupancy Rate,CT Lm:
Credit Tenant Lease,P SEm: Property Supply Elasticity,OEm,t: Office Using Employment in MSA,P ARm,t: Property Absorption Rate in MSA,F M Rt: Fixed Rate Mortgage Market Rate. The outcomes of M1a, M2a, M3a and M4a are default=0 and restructured=1. The outcomes of M1b, M2b, M3b and M4b are prepaid in cash=0 and defeased=1. The outcomes of M1c, M2c, M3c and M4c are curtailed=0 and partially defeased=1.
Figures in parentheses are standard errors.
***, ** and * represent significant level at 1%, 5%, 10%.
Table 5.10: Two-Stage Model (Winsorized): The Second Stage - Logit Model Results - Odd Ratios
Independent Variables M1a M1b M1c M2a M2b M2c M3a M3b M3c M4a M4b M4c
LT Vm,t 0.924*** 1.003 1.136*** 0.924*** 1.003 1.137*** 0.954*** 1.003* 1.136*** 0.930*** 1.003* 1.136*** T T Mm,t 1.0002 0.9993*** 0.998 1.0002 0.9993*** 0.998 1.000 0.9993*** 0.998 1.0001 0.9994*** 0.998 BP Rm,t 0.996*** 0.983*** 0.996*** 0.996*** 0.996*** 0.983*** 0.996*** 0.983*** N ARm,t 0.505*** 2.89×109* 0.505*** 3.27×109* 0.526*** 2.32×109* 0.511*** 2.67×109* M Rm,t 0.543*** 1.252*** 1.841*** 0.543*** 1.251*** 1.832*** 0.698*** 1.256*** 1.837*** 0.583*** 1.255*** 1.828*** M Tm 2.993*** 0.535*** 2.445*** 2.995*** 0.535*** 2.455*** 2.989*** 0.534*** 2.409*** 3.001*** 0.534*** 2.420*** CCm 1.222*** 1.537*** 1.222*** 1.538*** 1.219*** 1.537*** 1.228*** 1.538*** P P Y Mm,t 0.998 0.999 1.009 1.001 DSCRm,t 1.409*** 1.118** 0.223*** 1.408*** 1.117** 0.223*** 1.658*** 1.126** 0.230*** 1.466*** 1.125** 0.230*** P Am,t 1.005*** 1.006*** 1.022*** 1.005*** 1.006*** 1.021*** 1.005*** 1.006*** 1.022*** 1.005*** 1.006*** 1.021*** ORm,t 1.015*** 1.016*** 0.946*** 1.015*** 1.016*** 0.945*** 1.017*** 1.016*** 0.947*** 1.016*** 1.016*** 0.946*** CT Lm 0.293 44.766*** 0.294 44.626*** 0.321 44.732*** 0.359 44.599*** P SEm 0.257*** 6.276*** 1.72×10−6*** 0.258*** 6.324*** 1.37×10−6*** 0.316*** 6.214*** 2.1×10−6*** 0.275*** 6.260*** 1.69×10−6*** ∆P ARm,t−3 0.999 0.999 1.023 0.999 0.999 1.023 ∆F M Rt−3 0.992 3.212 323.425** 0.930 3.210 326.109** ∆IT Yt−3 1.001 0.978 0.888 1.006 0.978 0.888 Constant 632.359*** 0.002*** 0.005** 629.184*** 0.002*** 0.006** 1.833 0.049 34938 57.206*** 0.049 40242
Fixed State Effects Y Y Y Y Y Y Y Y Y Y Y Y
Fixed Time Effects Y Y Y Y Y Y Y Y Y Y Y Y
Notes:
LT Vm,t: LTV ratio,T T Mm,t: Time to Maturity,BP Rm,t: Balloon Payment Ratio,N ARm,t: Negative Amortization Rate,M Rm,t: Mortgage Rate,M Tm: Mortgage Type (Interest Only),
CCm: Cross Collateralization,P P Y Mm,t: Prepayment Penalty & Yield Maintenance,DSCRm,t: Debt Service Coverage Ratio,P Am,t: Property Age,ORm,t: Occupancy Rate,CT Lm:
Credit Tenant Lease,P SEm: Property Supply Elasticity,OEm,t: Office Using Employment in MSA,P ARm,t: Property Absorption Rate in MSA,F M Rt: Fixed Rate Mortgage Market Rate. The outcomes of M1a, M2a, M3a and M4a are default=0 and restructured=1. The outcomes of M1b, M2b, M3b and M4b are prepaid in cash=0 and defeased=1. The outcomes of M1c, M2c, M3c and M4c are curtailed=0 and partially defeased=1.
***, ** and * represent significant level at 1%, 5%, 10%.
As far as the decision between prepayment in cash and defeasance (model M4b) is concerned, a 1% change in mortgage rates, interest-only mortgages and the presence of cross-collateralization respectively determine an increase of 26% and reduction of 47% and 54% in the likelihood of defeasance. Similar results are found for this second pair of competing options for collateral terms: a 1% in- crease in DSCR and the existence of an underlying credit tenant lease increases defeasance occurrence by respectively 13% and more than 40 times.
Looking at property supply constraints, one unit increase in elasticity raises the probability of defeasance by five times, showing that mortgages are more likely and significantly to be defeased than prepaid in cash if the collateral is an office property in supply elastic markets. In fact lenders more likely offer defeasance provisions for mortgages collateralized in supply elastic market as property prices remain reasonably stable and the risk differential between holding Treasury se- curities and properties is reduced significantly.
In the second stage model, capturing changes in property absorption rates and market interest rates as well as Treasury yield does not bring any statistically sig- nificant changes in estimated results. Therefore we conclude that property supply factors are more important than demand side factors in explaining behaviour of early mortgage terminations. Our results confirm the research hypothesis and suggest the use of relaxation vs. tightening property supply constraints as a further risk management tool to control the exercise of mortgage termination options.