1 Data

## Load data 
load("~/Research/MMC_September_2020/hdata.houston.all.Rdata") # data.houston; all files 
rm(data.houston46)
load("df_hou_5.Rdata") # df_hou.s for structural analysis

# Load functions
source("MMC_functions_2020.R")

2 Descriptive Statistics

# Data set: data.houston 
summary.tab <- get.summary(dat= data.houston, 
                           var.fld = c("adr", "rating", "hi.sales", "room", "avmmc",
                                       "avmmc2"))

kable(summary.tab)

	M	Mean	Std.Err	Median	Min.	Max
adr	1880	85.8487	52.0007	68.3050	16.9900	400.75
rating	1880	1.7628	1.6594	1.0000	0.0000	6.00
hi.sales	1880	0.1376	0.1191	0.1003	0.0379	1.00
room	1880	111.6904	117.6942	80.0000	6.0000	1200.00
avmmc	1880	26.6886	29.2786	18.4714	0.0000	132.00
avmmc2	1880	1.1923	1.3344	0.8062	0.0000	9.00

3 Models for Market Definition

# functions
# get.avmmc
# get.avmmc.knn
# get.avmmc.cluster

3.1 Reduced Form

3.1.1 Models with AVMMC

# data set: data.houston
# OLS 
data.hou <- data.houston[data.houston$id.brand>1, ]
avmmc.ols0 <- lm(adr ~ avmmc, data= data.hou)
avmmc.ols1 <- lm(adr ~ avmmc + rating, data= data.hou)
avmmc.ols2 <- lm(adr ~ avmmc + rating + hi.sales, data= data.hou)
avmmc.ols3 <- lm(adr ~ avmmc + factor(chain), data= data.hou)
avmmc.ols4 <- lm(adr ~ avmmc + rating + factor(chain), data= data.hou)
avmmc.ols5 <- lm(adr ~ avmmc + rating + hi.sales + factor(chain), data= data.hou)


stargazer(avmmc.ols0, avmmc.ols1, avmmc.ols2, avmmc.ols3, avmmc.ols4, avmmc.ols5,
          type="html", omit =c("chain"), 
          add.lines = list(c("Chain FE", rep("No", 3), rep("Yes", 3))),
          title =  c("Results of OLS Models with AVMMC")
          )

**Results of OLS Models with AVMMC**

	Dependent variable:

	adr
	(1)	(2)	(3)	(4)	(5)	(6)

avmmc	0.059	0.089^***	0.088^***	-0.149^***	-0.064^**	-0.068^**
	(0.049)	(0.029)	(0.029)	(0.037)	(0.032)	(0.032)

rating		28.886^***	28.679^***		24.046^***	24.054^***
		(0.583)	(0.580)		(1.135)	(1.122)

hi.sales			-29.174^***			-33.394^***
			(6.679)			(6.099)

Constant	97.336^***	21.011^***	25.308^***	143.851^***	48.595^***	52.545^***
	(2.380)	(2.074)	(2.282)	(2.782)	(5.089)	(5.083)


Chain FE	No	No	No	Yes	Yes	Yes
Observations	1,274	1,274	1,274	1,274	1,274	1,274
R²	0.001	0.660	0.665	0.635	0.732	0.738
Adjusted R²	0.0003	0.659	0.664	0.628	0.726	0.733
Residual Std. Error	48.810 (df = 1272)	28.503 (df = 1271)	28.303 (df = 1270)	29.767 (df = 1249)	25.536 (df = 1248)	25.245 (df = 1247)
F Statistic	1.426 (df = 1; 1272)	1,231.605^*** (df = 2; 1271)	839.110^*** (df = 3; 1270)	90.625^*** (df = 24; 1249)	136.175^*** (df = 25; 1248)	135.130^*** (df = 26; 1247)

Note:	p<0.1; p<0.05; p<0.01

avmmc2.ols0 <- lm(adr ~ avmmc2, data= data.hou)
avmmc2.ols1 <- lm(adr ~ avmmc2 + rating, data= data.hou)
avmmc2.ols2 <- lm(adr ~ avmmc2 + rating + hi.sales, data= data.hou)
avmmc2.ols3 <- lm(adr ~ avmmc2 + factor(chain), data= data.hou)
avmmc2.ols4 <- lm(adr ~ avmmc2 + rating + factor(chain), data= data.hou)
avmmc2.ols5 <- lm(adr ~ avmmc2 + rating + hi.sales + factor(chain), data= data.hou)


stargazer(avmmc2.ols0, avmmc2.ols1, avmmc2.ols2, avmmc2.ols3, avmmc2.ols4, avmmc2.ols5,
          type="html", omit =c("chain"), 
          add.lines = list(c("Chain FE", rep("No", 3), rep("Yes", 3))),
          title =  c("Results of OLS Models with AVMMC2")
          )

**Results of OLS Models with AVMMC2**

	Dependent variable:

	adr
	(1)	(2)	(3)	(4)	(5)	(6)

avmmc2	0.520	2.593^***	2.788^***	-2.466^***	-0.451	-0.213
	(1.072)	(0.625)	(0.622)	(0.847)	(0.731)	(0.724)

rating		29.008^***	28.799^***		24.236^***	24.311^***
		(0.582)	(0.579)		(1.137)	(1.124)

hi.sales			-31.318^***			-32.949^***
			(6.666)			(6.120)

Constant	98.746^***	19.642^***	23.842^***	141.441^***	45.586^***	48.504^***
	(2.330)	(2.087)	(2.255)	(2.708)	(5.059)	(5.033)


Chain FE	No	No	No	Yes	Yes	Yes
Observations	1,274	1,274	1,274	1,274	1,274	1,274
R²	0.0002	0.662	0.667	0.633	0.731	0.737
Adjusted R²	-0.001	0.661	0.667	0.626	0.726	0.732
Residual Std. Error	48.833 (df = 1272)	28.418 (df = 1271)	28.186 (df = 1270)	29.856 (df = 1249)	25.572 (df = 1248)	25.290 (df = 1247)
F Statistic	0.236 (df = 1; 1272)	1,242.781^*** (df = 2; 1271)	849.615^*** (df = 3; 1270)	89.775^*** (df = 24; 1249)	135.652^*** (df = 25; 1248)	134.475^*** (df = 26; 1247)

Note:	p<0.1; p<0.05; p<0.01

3.1.2 Reduced Form with AVMMC.KNN

# starting data set: data.houston
avmmc.knn <- get.avmmc.knn(mkt.id.fld = "qtr", id.brand.fld = "id.brand", 
                           id.chain.fld = "id.chain" , price.fld = "adr", 
                           ocu.fld ="ocu", room.fld = "room", 
                           rating.fld = "rating",lat.fld = "lat", 
                           lon.fld = "lon", knn = 4, data= data.houston)

data.houston1 <- data.houston
for (var in names(avmmc.knn)){
  # print(var)
  temp <- avmmc.knn[[var]]
  temp <- as.data.frame(temp)
  if(length(grep("knn", var))==1){
    colnames(temp) <- var
  } else{
    colnames(temp) <- paste0(var, ".knn")
  }
  data.houston1[,colnames(temp)] <- temp
}


# data frame: data.houston1
# data set: data.houston
# OLS 
data.hou1 <- data.houston1[data.houston1$id.brand>1, ]
avmmc.knn.ols0 <- lm(adr ~ avmmc.knn , data= data.hou1)
avmmc.knn.ols1 <- lm(adr ~ avmmc.knn + rating, data= data.hou1)
avmmc.knn.ols2 <- lm(adr ~ avmmc.knn + rating + hi.sales, data= data.hou1)
avmmc.knn.ols3 <- lm(adr ~ avmmc.knn + factor(chain), data= data.hou1)
avmmc.knn.ols4 <- lm(adr ~ avmmc.knn + rating + factor(chain), data= data.hou1)
avmmc.knn.ols5 <- lm(adr ~ avmmc.knn + rating + hi.sales + factor(chain), data= data.hou1)


stargazer(avmmc.knn.ols0, avmmc.knn.ols1, avmmc.knn.ols2, avmmc.knn.ols3, avmmc.knn.ols4, avmmc.knn.ols5,
          type="html", omit =c("chain"), 
          add.lines = list(c("Chain FE", rep("No", 3), rep("Yes", 3))),
          title =  c("Results of OLS Models with AVMMC.KNN")
          )

**Results of OLS Models with AVMMC.KNN**

	Dependent variable:

	adr
	(1)	(2)	(3)	(4)	(5)	(6)

avmmc.knn	-1.574^***	-0.085	0.010	-0.967^***	-0.516^***	-0.419^***
	(0.163)	(0.104)	(0.105)	(0.122)	(0.108)	(0.109)

rating		28.693^***	28.657^***		23.207^***	23.442^***
		(0.614)	(0.610)		(1.142)	(1.134)

hi.sales			-29.459^***			-28.472^***
			(6.856)			(6.190)

Constant	109.343^***	25.547^***	28.816^***	140.586^***	50.835^***	52.712^***
	(1.656)	(2.056)	(2.179)	(2.071)	(4.767)	(4.746)


Chain FE	No	No	No	Yes	Yes	Yes
Observations	1,274	1,274	1,274	1,274	1,274	1,274
R²	0.069	0.657	0.662	0.648	0.736	0.740
Adjusted R²	0.068	0.657	0.661	0.641	0.730	0.735
Residual Std. Error	47.132 (df = 1272)	28.602 (df = 1271)	28.408 (df = 1270)	29.231 (df = 1249)	25.346 (df = 1248)	25.143 (df = 1247)
F Statistic	93.696^*** (df = 1; 1272)	1,218.694^*** (df = 2; 1271)	829.779^*** (df = 3; 1270)	95.906^*** (df = 24; 1249)	138.984^*** (df = 25; 1248)	136.611^*** (df = 26; 1247)

Note:	p<0.1; p<0.05; p<0.01

avmmc.knn2.ols0 <- lm(adr ~ avmmc.knn2 , data= data.hou1)
avmmc.knn2.ols1 <- lm(adr ~ avmmc.knn2 + rating, data= data.hou1)
avmmc.knn2.ols2 <- lm(adr ~ avmmc.knn2 + rating + hi.sales.knn, data= data.hou1)
avmmc.knn2.ols3 <- lm(adr ~ avmmc.knn2 + factor(chain), data= data.hou1)
avmmc.knn2.ols4 <- lm(adr ~ avmmc.knn2 + rating + factor(chain), data= data.hou1)
avmmc.knn2.ols5 <- lm(adr ~ avmmc.knn2 + rating + hi.sales.knn + factor(chain), data= data.hou1)

stargazer(avmmc.knn2.ols0, avmmc.knn2.ols1, avmmc.knn2.ols2, avmmc.knn2.ols3, avmmc.knn2.ols4, avmmc.knn2.ols5,
          type="html", omit =c("chain"), 
          add.lines = list(c("Chain FE", rep("No", 3), rep("Yes", 3))),
          title =  c("Results of OLS Models with AVMMC.KNN2")
          )

**Results of OLS Models with AVMMC.KNN2**

	Dependent variable:

	adr
	(1)	(2)	(3)	(4)	(5)	(6)

avmmc.knn2	-6.265^***	-0.414	-0.380	-4.093^***	-2.286^***	-2.276^***
	(0.680)	(0.430)	(0.432)	(0.508)	(0.449)	(0.449)

rating		28.677^***	28.731^***		23.180^***	23.295^***
		(0.611)	(0.614)		(1.138)	(1.140)

hi.sales.knn			-9.189			-15.577
			(10.883)			(9.877)

Constant	108.144^***	25.626^***	28.073^***	140.050^***	50.747^***	54.699^***
	(1.613)	(2.010)	(3.527)	(2.051)	(4.733)	(5.353)


Chain FE	No	No	No	Yes	Yes	Yes
Observations	1,274	1,274	1,274	1,274	1,274	1,274
R²	0.063	0.657	0.658	0.649	0.736	0.737
Adjusted R²	0.062	0.657	0.657	0.642	0.731	0.731
Residual Std. Error	47.285 (df = 1272)	28.600 (df = 1271)	28.603 (df = 1270)	29.206 (df = 1249)	25.315 (df = 1248)	25.299 (df = 1247)
F Statistic	84.910^*** (df = 1; 1272)	1,219.067^*** (df = 2; 1271)	812.765^*** (df = 3; 1270)	96.150^*** (df = 24; 1249)	139.450^*** (df = 25; 1248)	134.342^*** (df = 26; 1247)

Note:	p<0.1; p<0.05; p<0.01

3.1.3 Reduced Form Model with AVMMC.CL

\(AVMMC.CL_j\) is the average measure of \(MMC_{jk} = \frac{\sum_m I_j^m*I_k^m}{\sum_m I_j}\) over all firms in market \(m\).

# load new data set with variable cluster
load("~/Research/MMC_September_2020/data.houston.Rdata")

kable(table(data.houston$cluster), col.names = c("Cluster", "No. of Obs."),
      cpation ="Clusters in Houston, TX")

Cluster	No. of Obs.
0	233
1	458
2	224
3	57
4	118
5	23
6	65
7	156
8	56
9	134
10	16
11	23
12	20
13	56
14	33
15	16
16	16
17	60
18	16
19	56
20	16
21	28

data.hou.cl <- data.houston[data.houston$cluster>0,]
data.hou.cl <- data.hou.cl[order(data.hou.cl$qtr, data.hou.cl$cluster),]
mmc.cl <- get.avmmc.cluster(mkt.id.fld ="id.msa.qtr",id.brand.fld="id.brand",                             id.chain.fld ="id.chain", price.fld ="adr", 
                            ocu.fld = "ocu", room.fld ="room", 
                            revpar.fld = "revpar", rating.fld ="rating",
                            cluster.fld ="cluster", data = data.hou.cl)

data.hou.cl1 <- data.hou.cl
mmc.cl$avmmc <- unlist(mmc.cl$avmmc)
for (var in names(mmc.cl)){
  # print(var)
  temp <- mmc.cl[[var]]
  temp <- as.data.frame(temp)
  if(length(grep("cl", var))==1){
    colnames(temp) <- var
  } else{
    colnames(temp) <- paste0(var, ".cl")
  }
  data.hou.cl1[,colnames(temp)] <- temp
}

data.hou.cl0 <- data.hou.cl1
data.hou.cl1 <- data.hou.cl0[data.hou.cl0$id.brand>1,]

avmmc.cl.ols0 <- lm(adr ~ avmmc.cl , data= data.hou.cl1)
avmmc.cl.ols1 <- lm(adr ~ avmmc.cl + rating, data= data.hou.cl1)
avmmc.cl.ols2 <- lm(adr ~ avmmc.cl + rating + hi.sales.cl, data= data.hou.cl1)
avmmc.cl.ols3 <- lm(adr ~ avmmc.cl + factor(chain), data= data.hou.cl1)
avmmc.cl.ols4 <- lm(adr ~ avmmc.cl + rating + factor(chain), data= data.hou.cl1)
avmmc.cl.ols5 <- lm(adr ~ avmmc.cl + rating + hi.sales.cl + factor(chain), data= data.hou.cl1)


stargazer(avmmc.cl.ols0, avmmc.cl.ols1, avmmc.cl.ols2, avmmc.cl.ols3, avmmc.cl.ols4, avmmc.cl.ols5,
          type="html", omit =c("chain"), 
          add.lines = list(c("Chain FE", rep("No", 3), rep("Yes", 3))),
          title =  c("Results of OLS Models with AVMMC.CL")
          )

**Results of OLS Models with AVMMC.CL**

	Dependent variable:

	adr
	(1)	(2)	(3)	(4)	(5)	(6)

avmmc.cl	4.374^**	7.395^***	7.401^***	3.133^**	5.015^***	5.007^***
	(2.077)	(1.207)	(1.207)	(1.514)	(1.284)	(1.284)

rating		29.164^***	29.154^***		25.248^***	25.234^***
		(0.614)	(0.614)		(1.196)	(1.196)

hi.sales.cl			-12.229			-13.234
			(13.216)			(12.043)

Constant	97.447^***	16.918^***	17.997^***	131.937^***	34.531^***	35.754^***
	(2.550)	(2.249)	(2.534)	(2.835)	(5.201)	(5.318)


Chain FE	No	No	No	Yes	Yes	Yes
Observations	1,148	1,148	1,148	1,148	1,148	1,148
R²	0.004	0.665	0.665	0.622	0.729	0.730
Adjusted R²	0.003	0.664	0.664	0.614	0.723	0.723
Residual Std. Error	49.272 (df = 1146)	28.589 (df = 1145)	28.591 (df = 1144)	30.662 (df = 1123)	25.952 (df = 1122)	25.950 (df = 1121)
F Statistic	4.436^** (df = 1; 1146)	1,136.053^*** (df = 2; 1145)	757.559^*** (df = 3; 1144)	76.986^*** (df = 24; 1123)	120.990^*** (df = 25; 1122)	116.404^*** (df = 26; 1121)

Note:	p<0.1; p<0.05; p<0.01

Multimarket Contact in the Hotel Industry (Data Analysis)

Jung Hwan Koh

9/6/2020

1 Data

2 Descriptive Statistics

3 Models for Market Definition

3.1 Reduced Form

3.1.1 Models with AVMMC

3.1.2 Reduced Form with AVMMC.KNN

3.1.3 Reduced Form Model with AVMMC.CL

Cluster	No. of Obs.
0	233
1	458
2	224
3	57
4	118
5	23
6	65
7	156
8	56
9	134
10	16
11	23
12	20
13	56
14	33
15	16
16	16
17	60
18	16
19	56
20	16
21	28

Cluster	No. of Obs.
0	233
1	458
2	224
3	57
4	118
5	23
6	65
7	156
8	56
9	134
10	16
11	23
12	20
13	56
14	33
15	16
16	16
17	60
18	16
19	56
20	16
21	28

Cluster	No. of Obs.
0	233
1	458
2	224
3	57
4	118
5	23
6	65
7	156
8	56
9	134
10	16
11	23
12	20
13	56
14	33
15	16
16	16
17	60
18	16
19	56
20	16
21	28