Examination of Witnesses (Question Numbers
180-199)
MR RAPHAEL
WITTENBERG AND
PROFESSOR CAROL
JAGGER
5 NOVEMBER 2009
Q180 Dr Taylor: So it gets more complicated
by the moment, does it not?
Mr Wittenberg: Yes. In the model
in which I have been involved we try to look separately at different
groups and think about the overlap but yes.
Q181 Dr Taylor: Is there any way
of forecasting how much costs are going to go up?
Mr Wittenberg: Certainly we do
that although we do emphasise that we are not producing forecasts
in the sense that there is clearly no way of being sure what is
going to happen let alone people's expectations and wishes. What
we do is to produce what we call projections. That is to say that
we say "Supposing X, Y and Z, then how much would expenditure
be?" That is basically the approach we take. For example,
in the modelling which has been done for the Green Paper and earlier
modelling, we said, "Let's assume the numbers of older people
go up in line with the official ONS population projections. Let
us assume that the pattern of care remains the same, that the
probability of people with some particular characteristics receiving
care is constant. Let us also assume that the unit cost of care,
that is to say the cost of an hour's home care, rises by 2% a
year in real terms" or an alternative can be looked at. On
that basis, yes, we have produced projections and we produced
them originally for the Royal Commission for a whole lot of different
government agencies and the EU since then.
Q182 Dr Taylor: What is the margin
of error?
Mr Wittenberg: It is not possible
to give a statistical answer to that. It is not as though one
were just looking at, say, the confidence intervals of a particular
survey. Of course those exist, it is really that one has, with
each of these parameters we have been talking about, the numbers
of older people, disability, household composition, unit costs,
to think what the plausible range is in each. Then you either
vary them one after the other, which is what we have done, or
you take all the scenarios. We try to look at a minimum and a
maximum; the range is great and the biggest factor is the uncertainty
about unit costs because clearly our base case at the moment is
to assume unit costs rising about 2% a year in real terms but
that may be rather high.
Q183 Dr Taylor: We are coming on
to unit costs. One encouraging comment was from Professor Jagger
that some degree of obesity is protective in older people. As
somebody with a BMI above where it should be, I find that very
reassuring.
Professor Jagger: Overweight rather
than obese.
Q184 Sandra Gidley: I just want to
pick up on unit costs. You said that they would be increasing
2% a year in real terms. Can you clarify the reasons for saying
that? Why is it not just in line with inflation? Is it more people
or are there actual costs? No, if it is unit costs it is the cost
per person.
Mr Wittenberg: Actually by unit
cost, let us say of home care, I mean the cost per hour. Home
care is very, very labour intensive. In general terms—we
are not talking about the next couple of years but well ahead—the
Treasury have assumed that productivity and average earnings will
both rise by 2% a year in real terms. This is for very long-term
projections over the next 50 years and we have taken that as our
base case. Of course that may not be the case and there are lots
of reasons why it may not be. Particularly if one uses the 2%
assumption and particularly assuming that the average earnings
of care staff will go up in line with the average earnings of
the labour force generally, that may not be the case. I have seen
people arguing it both ways; some say care staff earnings might
rise faster and that could depend on future policy on the national
minimum wage, for example. So it is partly policy related. Another
view is that because the qualifications required are not very
high and there may be fewer jobs over time for people with lower
qualifications, the earnings of this group may rise more slowly.
There are also issues about expectations. Obviously if quality
of care rises, that may or may not affect the unit costs.
Q185 Sandra Gidley: Basically you
do not have a clue by the sound of it.
Mr Wittenberg: There is no categorical
answer. We can look at past trends and clearly the real unit costs,
particularly of care home provision, have risen but there may
be other reasons there as well. It is difficult. Sensitivity analysis
is wise and in the short term, given the present state of the
public finances, one might reasonably question whether 2% was
too high and perhaps the short term and the long term may differ.
Q186 Sandra Gidley: You also have
a demanding baby-boom generation coming up who may be demanding
improved services. Do you not think that is likely as well?
Mr Wittenberg: Yes and there has
been quite a lot of commentary on that and the general view, from
what I have read, is that is likely, although I have to say that
with home care that is more likely to affect the numbers of hours
per week perhaps than unit costs. With residential care of course
it could affect the weekly cost in terms of the facilities of
the care home.
Q187 Sandra Gidley: What is going
to be the impact of personalisation on the unit costs of long-term
care services?
Mr Wittenberg: There are two possible
issues here. Talking again about the cost of one hour's home care,
what one might suppose is, if block or bulk purchasing by councils,
some of whom have quite a bit of a monopsony power, that is to
say they are the only big purchaser in their area, that replacement
of that by lots of individual micropurchasing might give providers
the opportunity to raise prices through the diversification of
purchasing. There may be ways of handling that and it could be
that people could come together and purchase collectively or the
council could help them do that. There is a possibility of rising
unit costs from that source. Part of the evidence would be from
the care sector where it is well known that in general self-funders
seem to be charged more than the local authority pays. There is
one set of issues about unit costs in that sense of the word.
There is another set of issues about whether the weekly care costs
would be higher for people receiving personal budgets, let us
say, rather than people receiving conventional services and that
is something that the evaluation of the Ibsen project examined.
There is the second question as to whether the weekly care costs
will be different under individual budgets, now personal budgets,
from conventional packages of care. That is one of the issues
addressed in the evaluation of the individual budget pilots in
which the PSSRU and others were involved. They found no statistically
significant difference, as I understand it, in the average weekly
cost between the sample who were receiving individual budgets
and the control sample, the comparison with people who are not.
Q188 Sandra Gidley: Were the individual
budgets higher?
Mr Wittenberg: No, there was no
statistically significant difference.
Q189 Sandra Gidley: Sorry, I thought
you said they were higher.
Mr Wittenberg: There was no statistically
significant difference as I understand it.
Q190 Sandra Gidley: Would high rises
in unit costs impact on the suitability of the different funding
systems in the Green Paper?
Mr Wittenberg: I am just thinking
whether there is a link between those two matters. Personalisation
could apply on any of the funding systems in principle and therefore
the issue of whether personalisation would increase unit costs
and the issue of whether the funding mechanism would do something
are not the same. An issue which could arise for the comprehensive
scheme, which arose in Scotland with free personal care, is what
happens if the local authority were to seek to purchase for everybody,
the people who are now privately funded as well as those who are
publicly funded. My understanding of what happened in Scotland
is that care home providers were very unhappy about that and that
on the whole has not happened. Again, I suppose this could arise
under any of the funding systems so maybe the two issues are not
that closely related.
Sandra Gidley: So you do not think there
will be a connection.
Q191 Dr Stoate: When we talk about
people living on their own and having greater dependency needs
in the future, someone always talks about the dependency ratio
between the number of people economically active compared with
those economically inactive and how we sort that out. Is that
always the case? There are clearly younger people who are not
economically active and older people who are. What is your view
about whether that is a real issue or not?
Mr Wittenberg: The dependency
ratio I have come across is often simply the population ratio
of numbers of people over 65, over state pension age, and either
the rest of the population or the rest of the working-age population.
I think that people seem to find that, as far as I can tell, a
very useful, very quick summary indicator of the changing balance
between younger and older people. Clearly when one is looking
at long-term care demand one has to go rather deeper than that
and realise several things. One is that 40% of the gross expenditure
on social care actually relates to people below 65. Secondly,
of the 60% that relates to people above 65, the vast majority
actually relates to people above 75 and in many cases even older
than that. I find the dependency ratio perhaps very useful as
a headline starting figure but clearly in doing detailed work
we need more detail. In addition it has been pointed out that
there are plenty of people working beyond state pension age and
there are plenty of people who have retired before it.
Q192 Dr Stoate: Obviously productivity
tends to change over time, probably tends to increase in supply
over time and so as we go along outwardly improves. Does this
have implications for affordability, given that the country will
presumably become wealthier over time?
Mr Wittenberg: Yes. When we presented
our results from the model we have done we presented projected
expenditure relative to gross domestic product, national income
for that very reason. What complicates it, as we discussed earlier,
is the real rise in unit costs. There is in a way a link between
these things. The reason, for example, that the Treasury projects
in the long run—I am not talking about now of course—rises
in economic growth of GDP is partly issues of changes in the working-age
population, in the workforce, and partly increases in the per
person productivity and that in turn then feeds into issues about
average earnings and unit costs. Absolutely clearly one would
hope and expect that when we come out of the recession the economy
would grow again and that would affect affordability, but one
has to recognise that the same forces may cause the unit costs
to rise. The two are interrelated.
Q193 Stephen Hesford: Models and
projections and predictions and forecasts. Can you unpick what
they are? What should we really be talking about? Should we be
talking about models, predictions or projections? What do you
like to talk about?
Mr Wittenberg: I personally like
to talk about projections really because I feel that is what the
PSSRU, along with colleagues at Leicester and the University of
East Anglia and elsewhere, have been doing. By which I mean that
we are not saying this is what will definitely happen because
that, for reasons we have discussed, is impossible. We would rather
say "On this set of assumptions this is what we think expenditure
would be like. On that set of assumptions it will be something
else. This is what I would understand by projections. It is on
a set of assumptions and scenarios about some of the drivers of
expenditure. I personally talk about projections.
Q194 Stephen Hesford: In an historical
sense, I do not know how long you have been with the Department
or the Civil Service, maybe some years—
Mr Wittenberg: Yes.
Q195 Stephen Hesford: You have seen
a few projections and a few outcomes and the two may or may not
be the same thing and they may be vastly different. How confident,
based on your experience over time, have you become in your projections
as against what you know then actually happened?
Mr Wittenberg: That is something
we have sometimes been asked but never really pursued, to check
whether the projections produced for the Royal Commission met
with reality and we have not done this partly because we know
that so many factors have changed that we did not know about at
the time. It is doubtful that would be a useful exercise.
Q196 Stephen Hesford: From whose
point of view? Your sanity?
Mr Wittenberg: No, from the point
of view of commenting on whether the model is good or not. Because
if a lot has changed—
Q197 Stephen Hesford: Sorry, but
you said model now not projection.
Mr Wittenberg: Right. Maybe I
should say the model is the piece of software, for micro-simulation
it is a piece of computer code for the aggregate model is an Excel
spreadsheet. What is sitting on the computer is the model and
its output is the projection. I am sorry if I did not explain
that earlier.
Q198 Stephen Hesford: You do your
best and that is where we are.
Mr Wittenberg: Yes.
Q199 Stephen Hesford: You make the
assumptions and then unforeseeable variables which you have just
spoken about, changes in mortality, morbidity, fashion, risk factors,
medical progress, migration, inward, outward, birth rate all these
things. What is the margin? Do different variables have different
margins of error so that some are more confident than others or
are they all the same?
Mr Wittenberg: They are not all
the same. It is something we have looked at quite a bit because
it does depend how far ahead one makes the projections for; I
should make that point. The projections we have produced, particularly
to look ahead say 20 years or more, become particularly sensitive
to assumptions about unit costs; the 2% we were talking about
earlier versus 1% obviously compounds over a 20-year period and
that becomes enormous. By contrast, we have looked at life expectancy,
we have looked at the various official projections; ONS produce
low and high life expectancy projections and by contrast the difference
is considerable but it is not on that same scale. We did have
a bit of dialogue with ONS about that and they said that was a
function of what they felt were appropriate scenarios. Our colleague
at LSE, for example, Professor Mike Murphy, has produced a yet
higher life expectancy assumption and has explained at a recent
seminar his reasons for that, but even that is not certain. The
unit costs appear to be particularly important. Looking ahead
a long period of time again the dependency rates, the work we
have done with Carol Jagger, it is clearly very important that
over a long period of time, if there is compression or expansion
or morbidity, that matters. Informal care? We have found that
a bit less, if there were relatively small changes in the supply,
but again it depends hugely on assumptions one makes about what
would substitute for informal care. Similarly, if we get a change
in rates between residential and home care that will not make
a huge difference unless one believes that the home care package
is going to be a lot cheaper or a lot more expensive. My feeling
is that the biggest issues really are the unit costs and the disability
dependency issues in terms of sensitivity and perhaps I should
have said expectations as well which are of course almost unknowable.
| 
