Ippc

CONSUMER CHOICE AND

QUALITY OF HOME CARE

FOR THE ELDERLY

Mats Bergman, Economics, Södertörn University, Sweden

Henrik Jordahl, Research Institute of Industrial Economics, IFN, Sweden

Sofia Lundberg, Economics, Umeå University, Sweden

8/27/2014 1

List of care providers:

• Care for you Ltd

• Home quality

assistance Ltd

• Best care given Ltd

• Age with beauty

care Ltd

• …

AIM & MOTIVATION

• AIM: How does private provision and the introduction of consumer-

choice system influence quality, as perceived by the elderly?

• Provision of elderly care is the responsibility of the local government (municipal)

• (At least) three different systems:

1) In house production (public sector provision)

2) Procurement auctions (private provision)

3) Consumer choice (private provision)

8/27/2014 2

Privatization trend. Little is

known from the literature

Previous studies (Empirical) • Bedard & French, 2009, HeEc: Prison inmate mortality increases with medical-

staff contracting-out, by about 10 % • Bayer & Pozen, 2005, JLE: Recidivism among young prisoners higher after

release from private for-profit prisons, by about 10 % • Grabowski et al, 2008, HSR: Measurable quality in nursing homes do not fall

after conversion to for-profit status • Chou, 2002, JHE: Among information weak elderly, higher mortality in for-profit

nursing homes • Hsieh & Urquiola, 2006, JPuE: No effect of school voucher programs & private

schools on average educational outcomes • Bloom et al, 2010, WP: Increasing No of hospitals from 7 to 8 reduces heart-

attack mortality by (10 %?) • Gaynor et al, 2010, WP: Competition increases health care quality, without

increasing costs • Propper et al, 2008, EJ: Competition increases mortality in the UK (roughly by

10 %) • Gaynor, 2006, WP (survey): Most studies find competition increases hospital

quality when price is fixed • Bergman, Johansson, Lundberg & Spagnolo, 2014, (CEPR ,D P): Private

provision of nursing home care for the elderly: Decrease in mortality

8/27/2014 3

How?

• APPLICATION: Assisted living at home (basic medical

care, food services, hygiene assistance etc.)

• DATA: National survey with responses from approximately

70 000 elderly, in which the respondents were asked to

grade overall quality of the home-assistance service (from

2008)

• METHOD: Difference-in-difference at the municipal level

8/27/2014 4

We find evidence that consumer choice systems

increase perceived user satisfaction

Elderly care in Sweden

The State, the Central

Government (1)

The County (21)

The local government, the

municipality (290)

8/27/2014 5

1992

Systems, alternative ways to provide

elderly care

Elderly care

In-house Private In-house and

Private

The care

taker

No choice Consumer

choice

8/27/2014 6

In-house or private? In-house Private –

procurement

Private – consumer

choice

Responsibility

(by law)

Municipality Municipality Municipality

Provider In-house Private or mix Private or mix

Quality

monitoring

Municipality Municipality Municipality

How Call for tender – bids -

winner

Application – screening –

listing as provider

Senior

citizens

Apply & is

assigned a

provider

Apply & make a

choice or is assigned

a provider

Apply & make a choice.

If not happy? Switch.

Dominates

until 1990

Regulated since 1994 Regulated since 2009

Act of Free Choice

System (AFS)

8/27/2014 7

Three systems or two?

• Procurement or Act of Free Choice System?

• Senior residents perspective?

• Most likely equivalent, they have the opportunity to choose

• Quality outcome? Could be different. WHY?

Differences in:

• Competition – entry conditions

• Contract conditions (length of contracts)

• Screening?

• Quality standards more likely to be time consistent for Act of

Free Choice System

8/27/2014 8

Year of shift to Consumer Choice

8/27/2014 9

020

40

60

80

Fre

qu

ency

1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012Year of shift to Consumer Choice

Year of shift to Act of Free Choice System

8/27/2014 10

010

20

30

40

50

Fre

quency

2008 2009 2010 2011 2012 2013Year of shift to Act to Free Choice System

Illustration of data & research design

8/27/2014 11

2008 2013 Year

Procured

32

In-house

258

290

In-house

135

Procured

3

Consumer

Choice

(AFS)

152

3

135

Private

Data & method

• 290 municipalities

• 155 shift from in-house to consumer choice

• 152 shift from consumer choice by procurement to

consumer choice by the Act of Free Choice system

• A shift is not equivalent to 100 percent private provision

• Diff-in-diff on the municipal level using municipal fixed

effects

• Additional control: Cost per care-taker (annually)

8/27/2014 12

Cost per care taker

8/27/2014 13

0

20

04

00

60

0

2008 2009 2010 2011 2012 2013Year

Cost per care taker/1000 [price level 2008] Fitted values

Consumer satisfaction index (CSI)

• Measured annually. From 2008

• Survey to care takers: Overall impression (used in the regressions)

• Questions about specific services (ability to influence, opportunities for out-door activities)

• Answers: a grade indicating how satisfied the respondent is with overall impression and so forth

• Data 2008 – 2013

• Response rate about 70 percent, respondents #70 000 – 90 000

• Problem? Change in measurement method: 2012 and 2013

8/27/2014 14

The CSI measure

• Scale 0 – 100

• Prior to 2012:

• 1 – 10 scale and respondents indicated their overall impression of the care

• CISm = The average of all respondents in municipality m times 10

• From 2012:

• Six alternatives for overall impression:

• Very satisfied

• Pretty satisfied

• Neither satisfied or dissatisfied

• Pretty dissatisfied

• Very dissatisfied

• No opinion

8/27/2014 15

Share of

satisfied

respondents 0 –

100 = CSI

Illustration of the “problem”

8/27/2014 16

0

10

20

30

40

50

60

70

60

62

64

66

68

70

72

74

76

78

80

82

84

86

88

90

92

94

96

98

10

0

Fre

qu

en

cy

Index

2012

2011

Solution?

Percentiles (5 step intervals)

1 – 20

Support in the literature

Example

8/27/2014 17

10

12

14

16

18

CS

I P

erc

enti

le c

ate

gory

(1 -

20)

2008 2009 2010 2011 2012 2013Year

05

10

15

20

CS

I p

erc

en

tile

cate

go

ry (

1-2

0)

2009 2010 2011 2012 2013Year

Shift In-house

Municipal A Municipal B

Results

8/27/2014 18

Table 1. Results. Consumer satisfaction and shift to consumer choice

Dependent is

NKI-percentiles

Ordered probit Ordinary least square

VARIABLES Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

Shift to consumer 0.335*** 0.760*** 0.888*** 1.925*** 3.721*** 4.263***

choice (0.051) (0.082) (0.089) (0.321) (0.450) (0.483)

Cost per care

-0.002***

-0.008***

taker/1000 SEK

(0.001)

(0.003)

Constant

7.365*** 6.112** 6.513***

(0.192) (2.381) (2.379)

Municipal FE NO YES YES YES YES YES

Observations 1,740 1,740 1,713 1,740 1,740 1,713

chi2 42.93 662.7 676.8 . . .

LogL -5015 -4705 -4624 -5701 -5346 -5253

LogL0 -5036 -5036 -4963 -5719 -5719 -5630

Prob-value 5.68e-11 0 0 2.37e-09 0 0

F . . . 36.020 2.670 2.696

R2 . . . 0.020 0.348 0.356

R2 adjusted . . . 0.020 0.218 0.224

Standard errors in

parentheses

*** p<0.01, ** p<0.05, * p<0.1

Results

8/27/2014 19

Table 2. Results. Consumer satisfaction and shift to consumer choice and Act of Free

System (AFS)

Dependent is

NKI-percentiles Ordered probit

Ordinary least square

VARIABLES Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

Shift to -0.516*** -0.206 -0.147 -3.204*** -0.652 -0.362

consumer choice (0.134) (0.208) (0.209) (0.673) (0.939) (0.938)

Shift to AFS 0.949*** 0.979*** 1.067*** 5.735*** 4.418*** 4.750***

(0.137) (0.193) (0.195) (0.689) (0.832) (0.835)

Cost per care

taker 1000/SEK

-0.002***

-0.009***

Constant

(0.001)

(0.003)

7.359*** 6.803*** 7.334***

Municipal FE NO YES YES YES YES YES

Observations 1,740 1,740 1,713 1,740 1,740 1,713

chi2 91.27 688.9 707.6 . . .

LogL -4991 -4692 -4609 -5677 -5335 -5240

LogL0 -5036 -5036 -4963 -5719 -5719 -5630

Prob-value 0 0 0 0 0 0

F . . . 51.86 9.968 10.13

R2 . . . 0.047 0.357 0.365

R2 adjusted . . . 0.046 0.227 0.235

Standard errors in

parentheses

*** p<0.01, ** p<0.05, * p<0.1

Conclusions and extensions

• A shift to consumer choice by Act of Free System has a positive effect on consumer satisfaction

• A shift is equivalent to a 4 percentile step increase in CSI (20 percentiles).

• The idea of consumer power as a driver of quality might work!

• High costs are not equal to satisfied consumers …

Extensions

• Use the information on degree of competition (care takers with private provision as share of total number of care takers)

• Is cost endogenous?

8/27/2014 20