In a 1930s romantic comedy Greta Garbo plays an earnest Soviet commissar called Ninotchka dispatched to Paris to retrieve some Romanov jewels.
In a deadpan but memorable line she sums up Stalinism: "Fewer but better Russians."
Fewer but better teachers seems to be the Treasury's prescription for our school system.
Its briefing to the incoming Minister of Finance stirred up something of a wasps' nest by saying: "Increasing student/teacher ratios ... can free up funding that could be used to support initiatives to enhance the quality of teaching, such as more systemic use of value-add data and a more professionalised workforce."
One might hope, in passing, that better teaching would spare us such barbarous and murky language as "more systemic use of value-add data and a more professionalised workforce".
In any case, Treasury Secretary Gabriel Makhlouf felt it necessary to defend the comment, in a speech to the Trans-Tasman Business Circle on Tuesday.
"We've never said that class size doesn't matter. Far from it. But we are in a world where Governments, like households, have to make trade-offs and use their resources where they get the best results," he said.
"Very modest increases in class size, say on average one or two students per class across the system, would be unlikely to have a significant impact on achievement. Class size matters, but the quality of teaching matters more."
In a brief paper expanding on its advice the Treasury says Government spending per school student had increased by about 20 per cent in real terms over the past decade.
On average New Zealand kids do well in international surveys such as the OECD's PISA survey, which compares 15-year-olds in reading, mathematics and science.
But the average masks a substantial minority who do badly. Three out of 10 leave school without NCEA 2, including half of all Maori school-leavers.
Compared with most other developed countries, students' socio-economic background has a much bigger influence on achievement.
"Research on student learning consistently shows that the largest source of variation in student learning is attributable to differences in what students bring to school - their abilities and attitudes, and family and community background."
These are factors which are difficult for policy makers to influence, at least in the short-run, the Treasury says.
Makhlouf said it was a common misconception that low achievement was concentrated in particular schools or communities.
There was wide variation in achievement within almost all schools - the widest in the OECD, he said.
"I don't believe that three out of 10 students are simply too hard to teach, or are incapable of learning basic skills. The system is failing some students. To be sure, not all children are school-ready on their first day, and not all children come from families that are willing and able to support their learning," Makhlouf said.
"But then some of these students benefit from great teachers who inspire and guide, and help them achieve their potential.
"Others, however, get lost in the system, put in the too-hard basket, or fail to have specific learning needs identified and addressed."
The Treasury cites a recent report by Australia's Grattan Institute which highlights how four East Asian countries have achieved significant improvements in the performance and equity of their schooling systems by building teacher capacity.
"They have done so via a focus on high-quality initial teacher education, improved feedback and mentoring, and career structures that value good teaching."
In these respects, apparently, New Zealand's report card would read "Could do better."
Makhlouf stressed that they were not talking about crude performance pay or bonuses paid on the basis of test scores. "Rather, for example, a set of standards that distinguishes a developing teacher from a master teacher, with remuneration consistent with those roles. Measurement is hard, as always, but the egalitarian notion that all teachers are equally effective is simply not borne out in the evidence."
Data about how much students in a teacher's class have progressed over a year would be one input into assessing his or her performance.
But so would observations by other teachers or a principal of performance in class, and feedback from students and parents.
"I find it frankly incomprehensible that data on student achievement is seen as dangerous," Makhlouf said.
"Yes, data can be misinterpreted and misunderstood. But its value is immense and all organisations need data and information to help them improve."
On the question of class sizes the Treasury's survey of the literature encourages it to conclude that the evidence for an effect on educational achievement beyond the initial years of schooling is limited. "Little is known about the effects of class size on secondary-age students," it says.
But in almost the next breath it cites research which "suggests that the impact on student learning of moving from a class with an average teacher to one with a high-performing teacher is roughly equivalent to the effect of a 10-student decrease in class size".
Clearly those researchers did have a view on the impact of class size on educational attainment.
For an agency with a focus on the fiscal bottom line perhaps the most telling observation is that "the cost of reducing average teacher:pupil ratios by one student would require a 4 per cent increase in the number of teachers, at a cost of about $100 million a year".
The implication is that a rise in the ratio would deliver a similar saving.
The broader context in one of serious budgetary belt tightening.
The Government is committed to return the Budget to surplus by 2014/15.
The starting point is a deficit of $18 billion or 9 per cent of GDP last year. Half of that was the hopefully one-off cost of the Canterbury earthquakes, leaving a "fiscal effort" of about 4.5 per cent of GDP to be achieved over four years.
Education spending represents about one-sixth of core Crown expenditure. Vain to expect that it will escape the bean-counters.
Finance Minister Bill English, appearing before Parliament's finance and expenditure select committee yesterday, said the public sector represented about a quarter of the real economy and should be adding to productivity not subtracting from it.
The Budget Policy Statement has pencilled in an allowance for increased operating spending of $800 million a year over the next two years and $1.2 billion a year thereafter.
"That will permit cost increases in health and education, though in both cases we need to see what they can achieve with higher-than-ever investment in them," he said.
Beyond the highly contingent and precarious boost from some fraction of that allowance, the education budget is essentially frozen at about $12.2 billion a year, implying contraction in real terms.
But to recycle a slogan from the last bout of fiscal stringency: "Education cuts don't heal."