Introduction
The Republic of Cyprus, like most democratic governments, counted the welfare of citizens as an area of concern. Emulating a model that had been implemented for some time in America, the UK, and other European countries, the University of Cyprus and the Central Bank of Cyprus commenced a domestic version of the Survey of Consumer Finances in 1999. The study is so detailed – the authors report that respondents routinely needed an hour and a half to complete the questionnaire – that it has the potential to inform government policy extensively.
Discussion
Content and Findings
Karagrigoriou and Vonta (2006) report on the results of the triennial Cyprus Survey of Consumer Finances (CySCF) with the express purposes of:
- Defining the structure of family financial assets as they evolved between the 1999 and 2002 CySCF survey runs; and,
- Analyzing the propensity to go into debt.
The rationale for embarking on such analyses rests on the political, social, and economic importance of fielding the CySCF itself. As with all such data-gathering, importance is assessed from different perspectives that are either broad or specific in scope.
At the level of international economics, for example, the CySCF is rationalized on the basis that it also appears to be considered worthwhile elsewhere in Europe (Finland, Spain, Germany, England, the Netherlands) and America as well. Given that these all count among the wealthier industrialized economies, the island nation took some pride in being the only emerging economy to mount such a survey. From a national macroeconomic point of view, secondly, governments concerned with the welfare of their constituents also track economic wellbeing in point of income, savings, and other financial assets. Viewing matters from their perspective, political economists zero in on the debt burden that consumers bear for they are concerned with the financial risks that families expose themselves to.
In technical terms, this was a cross-sectional analysis of the 2002 benchmarking study. At the same time, the availability of the earlier study (the CySCF first ran in 1999) enables a tracking comparison akin to a two-interval longitudinal study.
The authors build on partial reporting of the CySCF since Antoniou et al. (2004, qtd. in Karagrigoriou and Vontae, op. cit.) had already done a longitudinal comparison of debt distribution while Kourouyiannis (2005, qtd. in Karagrigoriou and Vontae, op. cit.) had completed a cross-section profiling of Cypriot household assets form the initial survey run. Hence, Karagrigoriou and Vontae advance the state of knowledge by scrutinizing changes on both sides of the family balance sheet as well as reporting on changes in net wealth.
Taking the family as the unit under study, two external, macroeconomic factors came into play during the period under review. The first one, a bull run in the domestic stock market and then a crash, implies that families rode the bandwagon of speculative frenzy for realizing capital gains in listed stocks. At the end of 1999, the then-young stock market (established in early 1996) saw the general stock index rise a dizzying 688% on average volume more than 12 times greater than the previous year. That stock prices declined 65% in 2000 (as they inevitably do sooner or later) means that some capital gains were lost. From an econometric point of view, this effort is even more illuminating for quantifying the relationship between stock market activity and family incomes.
The second external development was fiscal, as the term is understood by economists to cover government policies and program implementation. As the 1999 CySCF commenced, the government embarked on a program of financial markets liberalization. The most important development, in the author’s view, occurred on 1st January 2001, when lending rate ceilings were abolished. This meant that banks and the mushrooming sector of non-bank financial institutions could charge whatever loan rates “the market could bear.”
Ordinary price theory does not explain the authors’ claim that the proportion of families offering some piece of property, other than their primary residence, as collateral rose by 50 percent (382). After all, if loans become pricier (interest rates rise markedly), rational consumers should be less prone to procure loans, whether collateralized by their primary residence or some other piece of property owned. One may infer then that other factors affected the propensity to borrow.
These other independent variables go unmentioned by the authors, presumably because they intended their desk research analysis primarily for domestic policymakers and they felt confident that such target readers shared some anecdotal evidence or other. Recalling the promotions and incentives that attend aggressive credit card marketing, one infers that the propensity for Cypriots to go into debt may well have been stimulated by liberal lending terms besides price.
This is a prime example of empirical research qua desk research on a large-base national sample survey. Given the focus on wealth creation, the national representative sample (539 respondents) was augmented with an over-sample of high-income households (558).
In data analysis, survey data were of course weighted to reflect the national (or population) distribution.
These are all standard fieldwork procedures.
The study instrument was the orthodox paper-and-pencil questionnaire mailed out to pre-listed households. Response rates for the general and high-income samples were not so different, “between 70% and 80%” respectively (p. 384), that any discussion of contribution to total sampling error is very likely inconsequential.
Since the authors go into detail about family ownership of financial instruments and wealth status, they needed to list choices in the questionnaire and define the various items in the balance sheet of a household:
“The ‘before-tax family income ‘is categorized into eight income categories according to the total amount of the yearly income of each household.
‘Assets” are categorized as financial and non-financial.
‘Financial assets’ cover liquid accounts like saving, deposit accounts, checking, government and saving bonds, mutual funds, saving certificates, stocks, corporate bonds, and warrants, the cash value of life insurance, other insurances like medical, educational, etc., cash value of pension plans and retirement accounts and liquid accounts in foreign currency.
‘Non-financial assets’ include consumer durables (like vehicles, motorcycles, etc), equity in businesses, the gross value of the primary residence, gross value of other residential estates, and gross value of the real estate in the part of the island which is not under the control of the Republic of Cyprus.
‘Pension plans’ are included [because the authors asset that] since we focus on …augmented wealth in which all sources of future income (e.g. pension plans) are included.
…[In 2002] retirement accounts …[were once again included in the definition of] social security contributions.
As to household liabilities, “types of household debts cover credit card balances, other lines of credit (store debit accounts), mortgage and home equity, loans for investment real estate, loans for businesses and education [and] “other loans” …[such as]: home improvement loans, student loans, consumer loans (overdraft facilities, household item loans, medical debts, loans from friends or relatives …), personal debts from businesses and loans against life insurance policies.” (p. 387)
Households were classified according to such family characteristics as family income, age and work status of the household head, education level, and housing.
“…a household consists of an economically dominant single individual or couple, all dependent children and any dependent adult living in the same house (e.g. elderly people, grandparents). The head of the household is defined as the economically dominant individual in the household.” (p. 388)
In general, the authors proved consistent in applying the technical terms they had defined.
Concerning valuation, the authors caution that claimed values for property in the area occupied by Turkey cannot be validated since the Republic has no regular access to, or proper knowledge of, the other half of the island.
Other than a somewhat conservative bias against ordinary consumers taking on debt, the authors must therefore constantly report real-world findings.
The highlights
The citizens of the Republic seem to be well-off if 99% ownership of real property and private vehicles is anything to go by. Awareness of local culture provides a partial explanation: parents are wont to gift children who settle down with either a house or substantial cash. Still, such beneficence does require a comfortable level of per-capita income.
Despite the sorry experience with the stock market downturn, the incidence of stock ownership was marginally higher (93%) in 2002 than in 1999 (89.8%).
The “trade-off” which the authors make much of, is that over the three years in question, the ownership rate for government savings bonds declined from 51% to 44%. The authors view this with disdain, in large part because their conservative economic paradigm makes safety/risk the prime consideration. They cannot understand how rational consumers can bear the risk of trading in the stock market when government bonds are so much safer. Karagrigoriou and Vontae need to be reminded that risk is inherent in all forms of entrepreneurship but this is counterbalanced by the prospect of immense financial rewards. And what is the stock market after all but a mechanism for marshaling equity capital from the public at large in return for having a voice in the affairs of a corporation?
The participation rate for mutual funds is low in the absolute (12.5%) and hardly better than the 10.2% observed in 1999. The authors explain this as owing to the lack of government encouragement or an attractive policy milieu.
Turning to the liability side of the family balance sheet, the authors report that close to two-thirds of Cypriots bore some debt or other and this proportion was unchanged from 1999.
However, the authors aver that the loan burden increased markedly among upper-income households. The value of mortgages on other real property comprised about one-sixth of aggregate household debt, rather higher than the figure of 13% discovered in 1999. Analysis of cross-breaks revealed that such loans represented fully eight in ten (Cyprus) pounds of the family’s total loans among those households residing in rented quarters. This stands to reason since renters very likely aspire for a home of their own.
Families’ income is highest when the household head is at the peak of his profession and earning power. Karagrigoriou and Vontae revealed that the mode of the distribution has shifted older, up to household heads between 50 and 59 years old. Otherwise, there were also palpable improvements in household income amongst families where the head was younger than 30 (p-value = 0.01) or where the household head was between 30 and 39 years (p-value = 0.07).
Across all categories, families were better off when the head was a government employee. This is explained by the fact that the Republic of Cyprus pay scale is very high.
More than two-thirds of families (70%) own their residence and another 13% rent to own.
For most households (80%) the median value of financial assets amounted to 4,520 pounds in 2002.
The stock market downturn notwithstanding, the proportion owning any stock increased markedly from 25.3% in 1999 to 36.5% by 2002 (statistically significant, p-value= 0.0001). That average holding declined in value (median of 5,350 Cyprus pounds in 1999 and 2,000 Cyprus pounds in 2002) presumably reflects the decline in stock values after the market fell back to more realistic prices in 2001.
Turning to “net wealth” (total assets less total liabilities), Karagrigoriou and Vontae resorted to graphical analysis. This revealed that beginning around the 25th percentile, families tended to have higher net wealth during the second run of the CySCF. This is presumptive proof for the “wealth mobility” of lower-middle, middle- and high-income families when macroeconomic fundaments of the country are sound.
On the other hand, debt accumulation was statistically higher than in 1999 starting with households in the 65th income percentile.
When the authors tested a binary, probit econometric model where the propensity to go into debt was expressed as a binomial dependent variable (have debt or not), Karagrigoriou and Vontae report that “as age increases the probability of one having a mortgage increases up to a certain point and then it starts to decrease as age keeps increasing.“ (p. 399) Stepwise regression went on to show that the peak probability for debt occurs when couples are in their thirties and early forties. This coincides with nearing or reaching the peak of their careers and earnings potential.
This is also the childrearing life stage: the family is growing and requires roomier accommodation.
Evaluation
Certainly, the task that the authors set themselves to is, by any objective measure, worth doing. In a political economy, it is good for a government to monitor the financial well-being of its citizens. That the United States and other leading economies had already implemented SCF before Cyprus took the initiative to do so also adds to the evaluation of its importance.
That said, inadequate analysis and unrealistic ideas from political economy taint this otherwise helpful report by Karagrigoriou and Vontae. The latter is the more persistent shortcoming.
Coming to what should be the straightforward treatment of objective survey data, the authors occasionally stumble in interpretation. Examples are:
Explaining stock holding that had reached 93% of the population purely based on an 85% prevalence of liquid accounts. The authors do not bother to explain the gap, forgetting that they had defined “liquid holdings” as including stocks in the first place. Hence, like some economists who get lost in complex models, they make the mistake of explaining B as an effect of A that also includes B.
The authors belabor an increase of 1.41 percentage points in the proportion of families who obtained loans against the real property other than their residence (4.96% in 1999 to 6.37% three years later). Even on a base of over one thousand returned questionnaires, no statistical test would yield a “significant” difference (i.e., greater than might be expected due to chance). The authors did not do any such test and based their conclusions exclusively on scrutinizing the basic summary tables. And yet, when analyzing the composition of assets and liabilities by household strata, the authors knew to use t-tests.
Stepping back, one must also realize that the absolute levels are in themselves really low. Other loan types, the mortgage on one’s residence, for instance) loom far large. So it may be artificial for the authors to point to loans on other real property as representing a significant shift in Cypriot behavior.
While one acknowledges that social and economic information is worth gathering only in so far as it informs government policy for public welfare, one must point out that the authors stray into side issues that do not necessarily strengthen the thrust of their analysis.
For instance, Karagrigoriou and Vontae complain that “the large amount of information collected in the Survey of Consumer Finances results in great size and demanding administration.” (p. 385) This is a specious argument because the University of Cyprus and the Central Bank of Cyprus had already taken care of data-gathering and tabulation. Hence, the authors merely had to analyze the available information.
Even were they to administer fieldwork and data entry themselves in a subsequent survey round, modern market research technologies cope just as well with the donkey work of 5,000 accomplished questionnaires as 500 or 50.
So, the authors waste printer ink proposing two other methods when they have no control over funding or the conduct of the study. One option they suggest is smaller specialized surveys. But a national survey database is valuable precisely for the variety of analyses it permits any interested third party. The notion of doing smaller surveys gains credence solely if one wishes policymakers to have the benefit of annual or semi-annual updates on the financial wellbeing of citizens. After all, a triennial run forces the assumption that change is linear over time.
Anybody who has ever tried to analyze in-depth movements in consumer incomes, spending, and savings realizes that these are dynamic phenomena, extremely sensitive and responsive to short-term market forces.
The authors also make much of the panel survey method in use in the Netherlands. This is a worthwhile point since panels remove one more source of sampling error.
On the other hand, the panel method is also susceptible to bias owing to the “Hawthorne effect”, defined as panel members being sensitized by the awareness that a third party regularly scrutinizes their spending and investment behavior. One rather suspects that financial matters are especially prone to this source of bias.
At the end of the day, neither the panel method nor periodic benchmarking research is inherently superior “for the evaluation of social indicators… factors responsible for economic inequality … mechanisms that can work effectively against it…issues of portfolio dynamics as well as the dynamics of living standards.” (p. 387)
Conclusion
At the end of the day, Karagrigoriou and Vontae completed the task they had set themselves to and that is to define the key findings of the 2002 CySCF across all information categories that the study comprised: assets, loans and liabilities, and net worth. Perhaps the work is tarnished by some oversights in data analysis and in not appreciating that the modern consumer does balance risk against potential rewards, besides going into some manageable debt to finance a desire for instant gratification and a worthwhile lifestyle. Most cope with repayment out of liquid assets and do not suffer grievously for it.
Bibliography
Karagrigoriou, A. (2005) The survey of consumer finances: Sampling and surveying in Cyprus, Luxembourg Wealth Study. Rome: Banca d’ Italia.
Karagrigoriou, A. & Vonta, I. (2006) On distributional changes of financial characteristics in Cyprus: What does the survey of consumer finances say? Financial Theory and Practice, 30 (4), 380-403
Kourouyiannis, K. (2005) The composition of Cyprus household assets: Results from the Cyprus Survey of Consumer Finances (CySCF1999). Diploma Thesis. Lefkosa: University of Cyprus, Dept. of Mathematics and Statistics.