Saturday, November 20, 2010

Is the CPI an appropriate basis for retirement planning?

With the risk of deflation likely avoided, more attention is now being given to the level of inflation. A number of countries have already started taking measures to curb excessive inflation (e.g. China, Australia). In simple terms, for consumers, if we experience inflation this means that the nominal cost of living will be higher in the future than it is today. If one assumes that there will be an inflationary economic environment going forward, that will have an impact on retirement planning and investment management. In order to address the issue, it is necessary to hazard a guess as to the amount of future inflation. Given that professional economists routinely get their estimates of near term inflation wrong, my first working assumption is that there is very little chance of my being able to accurately estimate future inflation rates over a longer time period, so I'll be honest with myself and call it a guess.

How much will obviously depend on where you live and how you spend you money. As an example, a family with school aged children and without employer subsidised health care will (most likely) face a larger cost of living increases than individuals and couples without school aged children and with employer subsidised health care. A family living in a developing economy will be more vulnerable to food price inflation than a family in a developed market. Location and individual circumstances matter. My own spending habits will also change materially over time. Accordingly, my second working assumption is that, even if CPI is an accurate measure of inflation (see below), I cannot rely on it as a measure of my own personal expenditure inflation.

Many countries construct a consumer price index to measure changes in the price of a selected basket of goods and services which consumers buy. Each country has its own basket/methodology and will periodically revise both the components in its basket and the weightings given to each component. While it is unarguable that consumer spending habits will change over time and that the CPI construction must be revised from time to time, there will be lags between changes in spending habits and the CPI basket being updated. My third working assumption is that CPI is always at least slightly outdated and, even if accurately constructed, will never accurately reflect what is happening in the real economy. Historically, there have been times when the CPI has badly lagged the real world - the energy shocks of the early 1970s being a classic example.

CPI numbers are constructed by governments. Governments have considerable incentive to understate the true rate of inflation, both for political and economic reasons. It has often been said that the Fed's job is to manage inflation expectations and beliefs, not to manage inflation itself. The decision to stop publishing M3 data is often cited in support of these claims. There are many people who argue that the understatement is very significant. The evidence (both for and against) is somewhat mixed. My own personal cost of living as certainly been rising much quicker than CPI. My fourth working assumption is that even if I don't subscribe to conspiracy theories, there is a very real possibility (probability?) that CPI is biased to understate the true rate of inflation.

My own personal basket of expenditures is heavily weighted to (i) housing (ii) education and (iii) travel. Post retirement, I will need to add (iv) health care to the list. My personal weighting far exceed the weightings given in the Hong Kong CPI basket. All of these items have historically increased at rates exceeding the general CPI. My fifth working assumption is that I should expect to experience a personal rate of inflation which is higher than the general rate of inflation.


The implications of my five working assumptions are sobering:

1. I need to assume that the real level of inflation I need to protect myself against will be higher than any official CPI data;

2. the required return on investments needs to be adjusted upwards to compensate for the expected level of inflation;

3. while my own personal rate of inflation is totally irrelevant to the pricing of assets in the marketplace, if inflation generally is understated, it follows that asset pricing and asset allocation needs to be revisited.

In short, CPI is not an appropriate surrogate for inflation in retirement planning. For me at least, I have to assume a higher rate of future increases in the cost of living.

Footnote: strictly speaking, inflation is a measure of the change in the money supply, not changes in the nominal prices of goods and services.

No comments: