There actually is an international standard for calibrating meters--ISO
2720-1974, but I don't know to what extent manufacturers follow it. The
equation that Rob mentioned is the basic calibration equation used in that
standard. As far as I know, Minolta and Pentax use K=14, and Sekonic use
K=12.5 (a difference of a sixth of a step). The standard calls for
calibration at 4700 K, but few manufacturers seem follow this
recommendation. I think most use a CIE A illuminant (2854 K); in any
event, many authorized service centers use a Kyoritsu or similar tester
with a 2854 K light source. I'm not sure the calibration color temperature
is the main issue, for reasons I discuss later.

Calibration to a given reflectance is a myth. A reflected-light meter is
calibrated by aiming it at a surface of known luminance, usually
transilluminated; there is no reflectance involved. It's possible to speak
of reflectance when comparing incident- and reflected-light meters. If you
refer to ISO 2720 and do the math, a reflected-light with K=14 and an
incident-light meter with C=250 (flat sensor) should agree when reading an
18% reflectance.

In practice, this isn't always case; the difference may arise partially
from specular reflections from a test card, and there may be some issues
with calibration, especially with incident-light meters. If incident-light
meters are calibrated according to ISO 2720, and one believes
manufacturers' stated values for C (typically, 250 for flat receptors, and
330-340 for hemispherical receptors), the meters should give noticeably
different readings when aimed at a point source with the different
receptors. Informal tests that I've done comparing several Minolta and
Sekonic meters don't show nearly the difference that I would expect.
Inquiries to Minolta and Sekonic got nowhere; in fact, no one seemed to
have the slightest idea of what I was talking about. I've pretty much
resigned myself to accept the discrepancies between the behavior I expect
and the behavior I see as one of life's mysteries ...

A far greater issue than minor differences in nominal calibration is the
differing spectral responses of different manufacturers' meters. Some
years ago, I noticed incredible differences (something 3-4 steps) between
an (unmodified) Pentax V and a Minolta Flashmeter III under low-pressure
sodium light. I had the responses of the two meters measured. The
difference was considerable: the Minolta was fairly narrow, resembling that
of the 1932 CIE standard observer, while that of the Pentax was quite
broad, with considerable sensitivity to UV and IR.

Admittedly, nearly monochromatic sodium light isn't typical photographic
illumination, but I've found that even if the two meters agreed at one
color temperature, they may not agree under different conditions.

It's outrageous that a thread like this ever should be necessary. I've
always been frustrated by the lack of agreement among different meters; a
good luminance photometer, such as a Minolta LS-100 or LS-110, has a
response very close to that of the CIE standard observer. Admittedly,
these meters are quite expensive, and the CIE observer isn't quite the
response that one would prefer for film, but I don't see why it should be
that difficult to get reasonable agreement to a standard that roughly
corresponds to that of a typical film. This indeed was one of the
objectives of the Zone VI-modified Pentax meter. I've never had the
response of my Zone VI Pentax digital measured, but it seems fairly close
to that of my unmodified Pentax V except when reading through orange or red
filters.

I discuss meter calibration more extensively at

http://www.largeformatphotography.info/articles/conrad-meter-cal.pdf